You are browsing the archive for fellowship.

Public Procurement Data in the Philippines and Where to Find It

- March 6, 2019 in fellowship, Reflections from the field

Ben Hur Pintor, our fellow from the class of 2018, here shares his thoughts and research on public procurement data in the Philippines. During the selection process for the 2018 School of Data Fellowship here in the Philippines, I was informed that the selected Fellow will be working with data related to public procurement. As I wasn’t a public procurement expert, I did a little research on the topic. Here, I’d like to share some of the interesting observations that I noticed:

Public Procurement Data in the Philippines

In theory, we expect public procurement in the Philippines to produce a lot of data considering how the process is defined by RA 9184 or the Government Procurement Reform Act. Under the law, public procurement includes all “acquisition of Goods, Consulting Services, and the contracting for Infrastructure Projects by any branch, department, office, agency, or instrumentality of the government” including procurement for projects that are wholly or partly funded by Foreign Loans or Grants pursuant to a Treaty or International or Executive Agreement unless different procurement procedures and guidelines are expressly stated or if the foreign loan and grant is classified as Official Development Assistance (ODA) under RA 8182 or the Official Development Assistance Act. From this definition alone, we can see that almost all government spending falls under public procurement and, thus, it is logical to assume that whenever the government spends, public procurement data should be produced. Aside from the definition of public procurement, the law also provides, as a general rule, that all procurement shall undergo Competitive Bidding except for specific cases when Alternative Methods of Procurement such as Limited Source Bidding, Direct Contracting, Repeat Order, Shopping, and Negotiated Procurement are allowed. These specific cases are subject to the prior approval of the Head of the Procuring Entity (HOPE) and should be justified by the conditions provided by the Act. Most of the time, Competitive Bidding which has the following steps — advertisement, pre-bid conference, eligibility screening of prospective bidders, receipt and opening of bids, evaluation of bids, post-qualification, and award of contract — is followed.

Steps in Public Procurement by Competitive Bidding

Each step in the public procurement process produces its own data — bid posts, pre-procurement and pre-bid conference proceedings, submitted bids, winning bids, information on the bidders, and the awarded contracts to name a few. There are also monitoring and evaluation documents and reports that are regularly created during the implementation of a government project and even after its completion. So with all this public procurement data supposedly being produced, where can it be found?

Where to Find It

The Government Procurement Reform Act or RA 9184 enacted in 2003 is the comprehensive law governing public procurement in the Philippines that put together all procurement rules and procedures covering all forms of government purchases from goods, to consulting, to infrastructure services. It sought to address the complexity and vagueness of public procurement and its susceptibility to abuse and corruption due to multiple procurement laws by simplifying and standardizing the procedures with a focus on transparency and accountability. The law added two interesting features to ensure transparency and accountability:
  1. the creation of an electronic portal which shall serve as the primary and definitive source of information on government procurement (PhilGEPS); and
  2. the establishment of the Government Procurement Policy Board (GPPB).
The PhilGEPS (Philippine Government Electronic Procurement System) is the country’s single, centralized electronic portal that serves as the primary and definitive source of information on government procurement. Government agencies, as well as suppliers, contractors, manufacturers, distributors and consultants, are mandated to register and use the system in the conduct of procurement of goods, civil works, and consulting services. On the website, the government can publish what goods, consulting services, and civil works projects it needs while suppliers, private contractors, and companies can search and view  these procurement opportunities. It features an Electronic Bulletin Board where all procurement opportunities, results of bidding, and related information are posted; a Registry of Manufacturers, Suppliers, Distributors, Contractors and Consultants; and an Electronic Catalogue of common and non-common use goods, supplies, materials and equipment. When fully implemented, the system is also intended have a Virtual Store, Electronic Payment System, and Electronic Bid Submission. The system is managed by the Procurement Service of the Department of Budget and Management.

The PhilGEPS website (version 1.5)

PhilGEPS also releases public procurement data published by different government agencies as mandated by the Government Procurement Reform Act together with other infographics and reports.

Some datasets available in PhilGEPS

 

Standard Reports and Datasets

 

Sample data (Number of Registered Organizations per Year)

Reports, Notices, and Infographics

  The GPPB, as established by the Government Procurement Reform Act, is an independent inter-agency body with private sector representation envisioned as the policy making entity and the governing body overseeing the implementation of procurement reform in the country. Its objectives include the preparation of a generic procurement manual and standard bidding forms for procurement; establishing a sustainable training program to develop the capacity of Government procurement officers and employees; and ensuring the conduct of regular procurement training programs by the procuring entities. It also stores and displays public procurement data submitted to it by procuring entities and regulatory bodies. These include information on Annual Procurement Plans, Procurement Monitoring Reports, List of Blacklisted Suppliers and Constructors, Constructors Performance Evaluation Summaries, Pre-Selected Suppliers and Consultants, List of Observers, and Status of Protests.

GPPB Website and Monitoring Data

Sample data (PDF format)

Aside from the PhilGEPS and GPPB, the different government agencies also publish procurement records on their respective websites in compliance with National Budget Circular No. 542. This Circular is more commonly known in the Philippines as the Transparency Seal Circular because it directs government agencies to have a Transparency Seal visible on their websites where the public can access information related to their agency. Some of the data that the circular requires to be released are: annual reports, approved budget and corresponding targets, major programs and projects, program and project beneficiaries, status of implementation and program/project evaluation and/or assessment reports, annual procurement plans, contracts awarded, and the name of contractors/suppliers/consultants. For example, the Department of Public Works and Highways has a Civil Works page on their website that shows key documents related to the public procurement of civil works projects.

DPWH Civil Works page

Is it Enough?

As highlighted by the Philippine Center for Investigative Journalism (PCIJ) on their report “Public Contracting in the Philippines: Breakthroughs and Barriers” about the infrastructure projects of the Department of Public Works and Highways (DPWH), there are challenges in terms of the completeness and accessibility of public procurement data in the country. Tracking the process from planning to implementation is difficult because not all the documents related to the procurement of infrastructure projects are published. This is compounded by the weak organization of files in agency websites which can confuse those unfamiliar with the procurement process. For example, even though the different documents related to one infrastructure project are available in the DPWH site, they are located on different web pages and are not linked to one another thus preventing users from easily understanding how documents might connect to each other. Aside from this, even though PhilGEPS and the GPPB are good sources of public procurement data, they are only repositories and are dependent on the data submitted to them by procuring entities. This becomes problematic when the procuring entities themselves fail or even refuse to submit their data. Another important thing I noticed about public procurement data in the Philippines is this: Publishing public procurement data in machine-readable formats is not (yet) the norm in the Philippines. If you look at the Government Procurement Reform Act, there is no mention about releasing or publishing procurement data and documents in machine-readable formats. The training programs by the GPPB designed to develop the capacity of procurement officers and employees for both the private sector and the national government agencies, government-owned and controlled corporations, etc do not include parts on working with or publishing machine-readable data. As a result, procuring entities and agencies release data without considering the implications of the format they are releasing it in. In fact, aside from those found in PhilGEPS, most of the public procurement data in the country are in non-machine-readable formats — as PDFs, documents, or even scanned images. Now, the procuring entities releasing the data might not consider this as a problem since compliance with the law only requires them to release the data but from the point of view of a data practitioner analysing public procurement data, a civil society organization creating visualizations in support of its advocacy, a journalist investigating government infrastructure projects, or even just a citizen trying to look for possible evidence of corruption in the procurement process, this adds a lot of extra steps to convert and standardize the data before any meaningful work can be done on it. Steps that could have been skipped had the data been released in a machine-readable format such as a spreadsheet, a comma-separated value  (CSV) file, or JavaScript Object Notation (JSON) file. One of the positive things pointed out by the PCIJ report was the opportunity to standardize, link, and publish more contracting data given by the current trend of government agencies creating or upgrading their information-management systems. This should be supported by efforts to raise awareness and convince the procuring entities, journalists, CSOs, and citizens of the benefits of releasing machine-readable data. Public procurement data should not be released just for the sake of releasing it. It should be released for the purpose of ensuring transparency, accountability, and equitability in the procurement process. To do this, it is imperative that the documents and information for each step in the procurement process, from planning to implementation, should be released in an open, transparent, and timely manner. Public procurement data should also serve the purpose of encouraging citizens, individuals, and organizations to keep themselves informed and engaged in how public money is spent. Towards this end, it is important to release data in formats such as spreadsheets, CSV, or JSON that make it easier for stakeholders to analyse, share, and re-use the data. One of the ways to ensure that data is easily shareable, analysable, and reusable is by following a standard like the Open Contracting Data Standard (OCDS). Of course, simply following a standard is not enough and could even be counterproductive when done without the right preparation. It is equally important to study how a standard complements the process and how it can be integrated with the current system.

Sources

Civil Works – Department of Public Works and Highways. http://www.dpwh.gov.ph/dpwh/business/procurement/civil_works/awarded_contracts Open Contracting Data Standard. Open Contracting Partnership. http://standard.open-contracting.org/latest/en/ Philippine Transparency Seal – Department of Budget and Management. https://www.dbm.gov.ph/index.php/about-us/philippine-transparency-seal Public Contracting in the Philippines: Breakthroughs and Barriers. Philippine Center for Investigative Journalism (PCIJ) with support from Hivos and Article 19. http://pcij.org/wp-content/uploads/2018/01/PCIJ.-Open-Contracting-in-Philippines-Report_01102018_b.pdf RA 9184 (Government Procurement Reform Act). https://www.gppb.gov.ph/laws/laws/RA_9184.pdf The 2016 Revised Implementing Rules and Regulations of RA 9184. https://www.gppb.gov.ph/laws/laws/RevisedIRR.RA9184.pdf The Government Procurement Policy Board. https://www.gppb.gov.ph/ The Philippine Government Electronic Procurement System. https://www.philgeps.gov.ph/ The Procurement Service. http://main.ps-philgeps.gov.ph/   Flattr this!

Using the procurement process as a lens for assessing audit reports: what to watch out for

- March 4, 2019 in fellowship, Reflections from the field

Odanga Madung, our 2018 Fellow, was fortunate to collaborate with the Institute of Economic Affairs in Kenya on their recent study into public procurement. In this article, Odanga reflects on his experiences and offers some tips for those tackling similar work.     The Institute of Economic Affairs (IEA) in Kenya recently carried out a study entitled ‘Public Procurement in Kenya: An Analysis of the Auditor General’s Reports’. I was fortunate enough to contribute as part of my fellowship with the School of Data. The Auditor General’s Office (OAG) was established in Kenya in 2004 under an Act of Parliament. Its aim is to provide independent oversight over how the Kenyan Government and its agencies spend taxpayers’ money. The audit process involves obtaining the accounts of a government entity, scrutinising them against proposed budget plans and contractual obligations, then providing a professional opinion on the state of the accounts. The OAG opinions consist of three types:
  • Unqualified: represents a clean bill of health. This means that the Auditor did not find any problem with the documentation and the entity has managed its funds properly.
  • Qualified: occurs when the Auditor General has found some problems but they are not pervasive. The auditor received all the information required for audit, but it revealed gaps in adherence to procedures and budgets.
  • Adverse: occurs when the auditor general is able to review the ministry’s documentation, but found pervasive problems and considerable changes will be necessary in order to rectify. This kind of finding should be of concern to oversight bodies.
  • Disclaimer: when the auditor is unable to review fully the ministry’s documentation because there is a substantial amount of information that the ministry has not made available. The record keeping is so bad that the auditor cannot give an opinion.
The IEA’s study looked at the Auditor General’s report through the lens of public procurement. They analysed the OAG’s reports by using the OC framework of the tender process, i.e. placing each violation in either the Pre Tender, Tender or Post Award stages. As a result, it highlighted what steps are often breached when the OAG does not give an unqualified opinion to a state entity’s accounts.   This was a much needed breath of fresh air in the corruption conversation that Kenyans are currently having. Mainly because it focussed on the how (the methods) rather than the what (numbers, figures and personalities) of corruption. I say this as corruption is not something that just happens, it is engineered.   The main finding of the study was that majority of procurement breaches tend to happen in the post award stage. A process that the IEA states often lends itself to the least public scrutiny and transparency in comparison to the other parts of the tender process in Kenya. This is very important in the Kenyan context because at the heart of the corruption problem in Kenya is the Tender process. However, very few Kenyans understand what it looks like. Few Kenyans also understand how the Tender process is used in the plunder of public funds. The reason the problems in the above paragraph exist are twofold:
  • Firstly, how the Kenyan media covers stories about corruption. They tend to focus on the figures lost and the personalities involved rather than how the money was stolen. This may be because media practitioners feel that is what will sell newspapers as opposed to producing reporting that may drive significant action both publicly and legislatively. It’s no surprise then that Kenya’s corruption coverage ends up echoing tabloid reporting. The fundamentals in understanding how corruption happens are missing at large.
  • Secondly, lack of public awareness on the intricacies of the tender process leads to lack of accountability demands from them. This is in part due to lack of government outreach and the current coverage afforded by the media.
  IEA’s report sought to address the problems above. The points below are some key lessons learned from my collaboration: It is important to define the professional opinions the Auditor gives and provide examples of what may lead to specific outcomes. The Auditor General’s report is a very technical document. The majority of ordinary citizens either tend to misunderstand or have no knowledge about the content at all. Given that they are a target audience for these reports, a key task when doing the research was to define the opinions that the auditor gives in a simple manner. Providing examples as to what each opinion meant was also important. Lack of a clear definition also lends itself to misinterpretation from the press, something that may lead to unintended consequences down the line.

Descriptions provided in the IEA report of the auditor opinions.

              For relatability, try to show how much expenditure each Opinion represents. This gives a clearer picture to audiences about how much of public spending comes under threat due to procurement violation in specific cases. Multiple levels of procurement breaches may occur and it may be worthwhile to highlight serial offenders. Corruption is something that is engineered to escape the prevailing systems of accountability in a country. IEA found that many procurement violations occurred at multiple stages of the tender process. In some cases they found unsupported expenditure leading to exaggerated prices for products, or single sourcing leading to incomplete projects that have been fully paid for. It is therefore important to highlight how many violations occur at multiple levels when carrying out such a study. Try your best to advocate for machine readability of report releases in machine readable formats to reduce errors that could be caused in transcribing. One of the biggest hurdles experienced in working with government reports currently in Kenya (and this would probably be the case in a lot of other African countries) is that the reports are produced in the form of scanned PDFs. It makes the process time consuming and error ridden due to transcription of the documents. This problem is something that we see being widespread across government institutions. As we press for better systems of accountability, making sure that accessibility of information is easier should be part of it. If you encounter such a problem, I would recommend using sandwich pdf (https://www.sandwichpdf.com/) to try to make majority of the text recognizable.

An example of one of the outputs from the Auditor General’s Office.

  Media houses have a habit of misinterpreting or exaggerating the findings found in such reports. Training them and holding them accountable for their reporting is important. Journalists and CSOs are a key conduit of this kind of information to the public. However, we have found cases where a lot of them do not understand the terminologies and reasoning contained in Audit reports. What this means is in an attempt to simplify the information for the public, a lot of it gets lost in translation. The IEA had an open forum with journalists explaining how to go about reading the report they wrote and the Auditor General’s as well. To conclude, the IEA did an amazing study that used the Open Contracting framework on the tender process to analyze the auditor general’s report. Corruption is a problem plaguing the developing world. However, audit and oversight organisations are gaining more powers and prominence in these countries. Looking at the information provided by them could reveal a lot about how corruption happens around the world. If you do decide to undertake a study like the one IEA did, the above points I mentioned should help you come up with a study that becomes an effective advocacy tool against corruption.   Flattr this!

“Not a scary concept”: Reflections from the Standard Group Data Conference

- October 10, 2018 in Event report, fellowship

In his first piece for the School of Data blog, our 2018 Fellow, Kelvin Wellington, reflects on his experiences at the Standard Group Conference in Accra in July 2018. To date, the conversation around open data has been firmly centred in its importance and the implications of championing the cause. Is it a cause worth fighting for, and are policymakers doing the right thing by opening up data to the public eye? As citizens, is it important to know the finer details of how our country is run? These are questions that were lingering in my mind during an open data presentation that was part of a data conference held by the Standard Group in Accra, Ghana a few days ago. I will attempt to dissect some findings from this session. What can open data do? Open data should not be a scary concept, and should be embraced. It should not be seen as a means of taking off ‘protective shields’ on data. When we talk about open data, we should be looking at the following:
  • Empowerment: open data can give citizens of a country a stronger voice on public services they use and create a channel of dialogue between the citizen and local authorities or government.
  • Transparency: open data should be the next frontier in citizens’ quest for transparency. Freedom of information enables citizens to make informed decisions regarding their government, and allows us to better understand our world.
  • Participation: open data should bring about inclusiveness; from data providers to users. Everyone has a part to play in innovating with data and making a difference through building data-driven solutions.
  Who should be driving open data? Ideally, policy makers should be the driving force for open data in any setting. Policy makers in Ghana are, however, driving at a turtle’s pace. The public should be weighing in on the conversation as well, but at the moment thoughts are too scattered to produce a collective force. The private sector should also be heavily involved in the Open Data push since they have access to huge amounts of data. In addition, the policies and procedures should be open as well, not just data. Ultimately, it is up to governments, public bodies, community groups, citizens and businesses to facilitate the growth of open data, propagate its benefits and see that it achieves its full potential. The Open Data Initiative in Ghana has stalled with the online platform lacking in up-to-date data, and data unavailable for a good number of industries. Financial constraints have been pointed out as a major issue, and as citizens, we owe it to our country to challenge authorities to resolve this. Why is it important? We spend our time talking about making decisions without focusing on making data-driven decisions. If data is not being processed into knowledge and that knowledge does not become wisdom, then the purpose of data in itself is dead. Opening data gives us all a chance to contribute to creating more knowledge and making wiser decisions. Data has become a gold standard, and keeping an ‘open culture’ makes for a healthier ecosystem for policymakers and citizens alike.     Flattr this!

Reflections from the Field #2: Cultural Heritage Conservation as a Data Problem

- October 8, 2018 in fellowship, Reflections from the field

For the second in our series of blog posts – ‘Reflections From The Field’ – our 2018 School of Data Fellow, Ben Hur Pintor, is inspired to ask questions about cultural heritage data. “Can this be considered a data problem?” asked Tatine, one of the participants of the Data Pipeline training conducted by Hani and I, referring to her work on the conservation of the Mangyan script and language of the Mangyan indigenous people in Mindoro, Philippines. The question piqued my interest because, as mentioned by Tatine, indigenous communities rarely have physical records of the information related to their culture and language. Most of the time, this information is passed from generation to generation as a form of oral tradition or oral history. So how do we take something intangible and make data work for it?

Fieldworks

One of the most common ways of working with language and other cultural heritage data is through the conduct of fieldworks. Terry Crowley’s Field Linguistics: A Beginner’s Guide highlights this process, discussing ethical issues such as informed consent and voluntary participation, the importance of selecting language helpers — people who speak the language you are studying but also share a common language with you, and how to keep track and archive the data using daily records, filing systems, and computer storage.

Archives

Creating digital archives is also one way to work with cultural heritage data. This usually involves scanning, transcribing, and digitizing artifacts, artworks including poetry and song, and even places such as sites and monuments. Conservation measures applicable to the physical heritage are not appropriate for the intangible heritage. It is necessary therefore to establish digital archives by recording these cultural expressions on both visual and audio media to facilitate their survival and transmission to future generations.“ (Outline of Digital Archiving Project, UNESCO)

Linked Open Cultural Heritage Data

Linked open data is open data available on the world wide web in a standard markup format. It has a lot of potential uses in cultural heritage conservation especially when utilized together with traditional cultural heritage institutions such as libraries, archives, and museums. With linked open data, users can create and share their own cultural heritage experiences using different media, applications, maps, etc. “We believe linked open data has the potential not just to preserve cultural heritage for users, but to offer users new opportunities to understand, manipulate, and recreate cultural heritage experiences.” (Linked Open Data for Cultural Heritage, J. Marden, 2013) The combination of data, technology, and cultural heritage is an interesting one and it will be fascinating to see how these three fields interact, grow, and collaborate with each other in the future. Flattr this!

Introducing the 2018 Class of School of Data Fellows!

- June 22, 2018 in announcement, fellowship

School of Data is delighted to announce its sixth class of fellows. From June until January 2019, the programme will allow fellows to deepen their data literacy skills and work alongside local partner organisations to enhance the data literacy network local to them. We were really pleased to receive a large number of applications and would like to both congratulate and wish all our new fellows the very best for their fellowship! Pamela Gonzales is passionate about data visualization and bridging the digital divide for women. She is the co-founder of Bolivia Tech Hub, a collaborative space for tech projects to contribute to the prosperity of an innovative ecosystem in Bolivia. Pamela is also the Regional Ambassador for Technovation, a San Francisco based program that equips girls with the skills needed to solve real-world problems through technology. She holds a Bachelor of Science degree in Computer Science from Universidad Mayor de San Andres.       Odanga Madung is the co-founder and Data Science Lead at Odipo Dev, a data science and analytics firm operating out of Nairobi Kenya that delivers services to various bluechip companies and NGOs across the country. Odanga’s deepest interest is at the intersection between data and culture and it is through this that Odipo Dev has been able to carry out data analysis and visualisation on various activities for a wide range of clients and occurrences in Kenya and the world.Some of his work has been featured in publications such as Adweek, Yahoo, BBC, CNBC, Quartz, and Daily Nation, just to mention a few. He will be working on Open Contracting in Kenya during the period of his fellowship. You can follow him on Twitter @Odangaring and Odipo Dev @OdipoDev for more information.   Nzumi Malendeja is a Research Associate at an Independent Evaluation and Research Cell of BRAC International in Tanzania, where he leads larger-scale research projects in education, agriculture, and health. Here, he has developed mobile-based data collection platforms (ODK Collect and SurveyCTO), which replaced the traditional paper-based methods. Before this, Mr. Nzumi worked as a Field Monitor and Research Assistant at SoChaGlobal and Maarifa ni Ufunguo respectively, both in education and construction sector transparency projects. Mr. Nzumi has attended a 4 week Summer School Training on Research Methods and Teaching Skills, hosted by Hamburg University of Applied Sciences in Germany, funded by the Germany Academic Exchange Services (DAAD). Presently, Mr. Nzumi is working on his thesis towards the fulfillment of the Master of Research and Public Policy at the University of Dar es Salaam.   Sofia Montenegro A fan of nature and the teachings it hides, Sofia has dedicated herself to research in the social sciences. She studied Political Science at the Universidad Francisco Marroquin and Public Opinion and Political Behavior through a Masters degree at the University of Essex, where she deepened her interest in data methodologies in social research. Sofia is interested in academia only as long as it drives political action. She looks to help other women to be involved freely in data practice and political spaces. Sofia is also interested in network analysis, studying corruption as a social phenomenon, following electoral processes and learning research methods.   Elias Mwakilama is a lecturer at University of Malawi-Chancellor College and Coordinator of Research, Seminar and Consultancies, and Diploma in Statistics programme in the Mathematical Sciences Department, Elias Mwakilama is a computational and applied mathematician in the field of operations research. He lectures and supervises undergraduate students in Mathematics & Statistics fields. His research interests are in working with optimisation models using mathematical statistics techniques integrated with computing skills to offer solutions of industrial related problems in theoretical and practical arena. Elias holds a first upper class MSc degree in Mathematical Sciences from University of Malawi. His website is here. During his fellowship, he hopes to support the “public procurement open contract platform” for Civil Society Organisations (CSOs) in Malawi with Hivos.   Ben Hur Pintor is an open-source and open-data advocate from the Philippines​ who believes in democratising not only data, but ​also ​the means of utilising and analysing data.​ He’s a geospatial generalist and software developer who’s​ worked on projects related to renewable energy, blue carbon ecosystems, and participatory disaster risk mapping and assessment. ​Ben is currently pursuing an MS Geomatics Engineering degree at the University of the Philippines. As part of his advocacy for Free and Open Source Software (FOSS), he’s a co-organiser and active participant of FOSS4G Philippines and MaptimeDiliman — avenues for sharing open​ ​source mapping technologies with the community.   Hani Rosidaini is passionate about how technology can be adopted and applied for people’s needs. She combines her technical skills, especially in information systems and data science, with social and business knowledge, to help companies and organisations in Indonesia, Australia, and Japan. This includes her own ventures. Highly relevant to this year fellowship’s focus of data procurement, Hani has experience as a data specialist for public policy in the Indonesia Presidential Office, where she has analysed the national integrated data platform, data.go.id, contributed to data-driven policy making, advocated ministries and agencies, as well as engaged with civic and local communities. Flattr this!

Introducing the 2018 Class of School of Data Fellows!

- June 22, 2018 in announcement, fellowship

School of Data is delighted to announce its sixth class of fellows. From June until January 2019, the programme will allow fellows to deepen their data literacy skills and work alongside local partner organisations to enhance the data literacy network local to them. We were really pleased to receive a large number of applications and would like to both congratulate and wish all our new fellows the very best for their fellowship! Pamela Gonzales is passionate about data visualization and bridging the digital divide for women. She is the co-founder of Bolivia Tech Hub, a collaborative space for tech projects to contribute to the prosperity of an innovative ecosystem in Bolivia. Pamela is also the Regional Ambassador for Technovation, a San Francisco based program that equips girls with the skills needed to solve real-world problems through technology. She holds a Bachelor of Science degree in Computer Science from Universidad Mayor de San Andres.       Odanga Madung is the co-founder and Data Science Lead at Odipo Dev, a data science and analytics firm operating out of Nairobi Kenya that delivers services to various bluechip companies and NGOs across the country. Odanga’s deepest interest is at the intersection between data and culture and it is through this that Odipo Dev has been able to carry out data analysis and visualisation on various activities for a wide range of clients and occurrences in Kenya and the world.Some of his work has been featured in publications such as Adweek, Yahoo, BBC, CNBC, Quartz, and Daily Nation, just to mention a few. He will be working on Open Contracting in Kenya during the period of his fellowship. You can follow him on Twitter @Odangaring and Odipo Dev @OdipoDev for more information.   Nzumi Malendeja is a Research Associate at an Independent Evaluation and Research Cell of BRAC International in Tanzania, where he leads larger-scale research projects in education, agriculture, and health. Here, he has developed mobile-based data collection platforms (ODK Collect and SurveyCTO), which replaced the traditional paper-based methods. Before this, Mr. Nzumi worked as a Field Monitor and Research Assistant at SoChaGlobal and Maarifa ni Ufunguo respectively, both in education and construction sector transparency projects. Mr. Nzumi has attended a 4 week Summer School Training on Research Methods and Teaching Skills, hosted by Hamburg University of Applied Sciences in Germany, funded by the Germany Academic Exchange Services (DAAD). Presently, Mr. Nzumi is working on his thesis towards the fulfillment of the Master of Research and Public Policy at the University of Dar es Salaam.   Sofia Montenegro A fan of nature and the teachings it hides, Sofia has dedicated herself to research in the social sciences. She studied Political Science at the Universidad Francisco Marroquin and Public Opinion and Political Behavior through a Masters degree at the University of Essex, where she deepened her interest in data methodologies in social research. Sofia is interested in academia only as long as it drives political action. She looks to help other women to be involved freely in data practice and political spaces. Sofia is also interested in network analysis, studying corruption as a social phenomenon, following electoral processes and learning research methods.   Elias Mwakilama is a lecturer at University of Malawi-Chancellor College and Coordinator of Research, Seminar and Consultancies, and Diploma in Statistics programme in the Mathematical Sciences Department, Elias Mwakilama is a computational and applied mathematician in the field of operations research. He lectures and supervises undergraduate students in Mathematics & Statistics fields. His research interests are in working with optimisation models using mathematical statistics techniques integrated with computing skills to offer solutions of industrial related problems in theoretical and practical arena. Elias holds a first upper class MSc degree in Mathematical Sciences from University of Malawi. His website is here. During his fellowship, he hopes to support the “public procurement open contract platform” for Civil Society Organisations (CSOs) in Malawi with Hivos.   Ben Hur Pintor is an open-source and open-data advocate from the Philippines​ who believes in democratising not only data, but ​also ​the means of utilising and analysing data.​ He’s a geospatial generalist and software developer who’s​ worked on projects related to renewable energy, blue carbon ecosystems, and participatory disaster risk mapping and assessment. ​Ben is currently pursuing an MS Geomatics Engineering degree at the University of the Philippines. As part of his advocacy for Free and Open Source Software (FOSS), he’s a co-organiser and active participant of FOSS4G Philippines and MaptimeDiliman — avenues for sharing open​ ​source mapping technologies with the community.   Hani Rosidaini is passionate about how technology can be adopted and applied for people’s needs. She combines her technical skills, especially in information systems and data science, with social and business knowledge, to help companies and organisations in Indonesia, Australia, and Japan. This includes her own ventures. Highly relevant to this year fellowship’s focus of data procurement, Hani has experience as a data specialist for public policy in the Indonesia Presidential Office, where she has analysed the national integrated data platform, data.go.id, contributed to data-driven policy making, advocated ministries and agencies, as well as engaged with civic and local communities.   Kelvin Wellington is a Data Scientist from Accra, Ghana. He holds a Masters degree in Data Science from the University of Southampton as well as a first degree in Computer Science from Ashesi University, Ghana. He has had the opportunity to work in various roles that involve the application of data-driven solutions. He is passionate about using data for social good and has been involved in various volunteering projects to that effect..
Kelvin is also an active member of the Machine Intelligence Institute of Africa(MIIA) and a facilitator for Data Science meetups organised by Developers in Vogue, a tech community based in Ghana. He will be joining the 2018 fellowship class to work with the Natural Resource Governance Institute on data literacy in the extractives sector.” Flattr this!

Apply Now! School of Data’s 2018 Fellowship Programme

- April 16, 2018 in announcement, bolivia, fellowship, ghana, Guatemala, indonesia, kenya, Malawi, philippines, tanzania

School of Data is inviting journalists, data scientists, civil society advocates and anyone interested in advancing data literacy to apply for its 2018 Fellowship Programme, which will run from May 2018 to January 2019. 8 positions are open, 1 in each of the following countries: Bolivia, Guatemala, Ghana, Indonesia, Kenya, Malawi, Tanzania, The Philippines. The application deadline is set on Sunday, May 6th of 2018. If you would like to sponsor a fellowship, please get in touch with School of Data at Apply for the Fellowship Programme

The Fellowship

Fellowships are nine-month placements with School of Data for data-literacy practitioners or enthusiasts. During this time, Fellows work alongside School of Data to build an individual programme that will make use of both the collective experience of School of Data’s network to help Fellows gain new skills, and the knowledge that Fellows bring along with them, be it about a topic, a community or specific data literacy challenges. Similarly to previous years, our aim with the Fellowship programme is to increase awareness of data literacy and build communities who together, can use data literacy skills to make the change they want to see in the world. The 2018 Fellowship will continue the work in the thematic approach pioneered by the 2016 class. As a result, we will be prioritising candidates who:
  • possess experience in, and enthusiasm for, a specific area of data literacy training

  • can demonstrate links with an organisation practising in this defined area and/or links with an established network operating in the field
We are looking for engaged individuals who already have in-depth knowledge of a given sector or specific skillsets that can be applied to this year’s focus topics.. This will help Fellows get off to a running start and achieve the most during their time with School of Data: nine months fly by! Read More about the Fellowship Programme

The areas of focus in 2018

We have partnered with Hivos and NRGI to work on the following themes: Procurement and data in the extractives industry (oil, mining, gas). These amazing partner organisations will provide Fellows with guidance, mentorship and expertise in their respective domains.

2018 Fellowship Positions

Bolivia The Fellowship in Bolivia will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: Experience with and interest in community building, experience with the implementation of civic projects with a data or technical component, storytelling skills, and experience with promoting data or technical stories to a wide audience, basic understanding of the public procurement process Guatemala The Fellowship in Guatemala will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: Experience in the planning, coordination and implementation of projects with civil society organisations, the ability to advise and train organisations on working with data and delivering technical projects, basic understanding of the public procurement process Ghana The Fellowship in Ghana with be focused on extractives Data through the Media Development Programme at NRGI. For this position, School of Data is looking for someone with: an interest in supporting or working within the civil society sector, experience working with financial (or related) data for analysis experience as a trainer and/or community builder, interest and/or experience in the extractives sector, demonstrated skills as a data storyteller or journalist Malawi The Fellowship in Malawi will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering technical and data-driven projects, experience with facilitating training activities, experience with data collection projects, basic understanding of the public procurement process **Indonesia ** The Fellowship in Indonesia will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering technical and data-driven projects, experience with facilitating training activities, experience with working with government systems or data. Candidates with the following optional interests and experience will be appreciated: experience with explaining complex topics to varied audiences, experience with user design methodologies, experience with community development The Philippines The Fellowship in The Philippines will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with user-centric research and design methodologies, experience with community-building activities, experience with data storytelling. Candidates with the following optional interests and experience will be appreciated: graphic design skills, experience with delivering trainings Kenya The Fellowship in Kenya will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering data-driven projects, experience with user research and data storytelling, experience with explaining complex topics to varied audiences. Candidates with the following optional interests and experience will be appreciated: interest in or experience with supporting civic projects and civil society organisations, experience with facilitating training activities. Tanzania The Fellowship in Tanzania will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering data-driven projects, experience with facilitating training activities, experience with explaining complex topics to varied audiences. Candidates with the following optional interests and experience will be appreciated: experience working with journalists or as a journalist, interest in or experience with supporting civic projects and civil society organisations, experience with writing pedagogical content 9 months to make an impact The two programmes will run from May to Jan uary 2019, and entail up to 10 days a month of time. While Fellows will be focused on ironing their skills as data trainers and build a community around them, Experts will focus on supporting and training a civil society organisation or newsroom with a specific project. Fellows will receive a monthly stipend of $1,000 USD a month to cover for their work. In May, both Experts and Fellows will come together during an in-person Fellowship Induction Workshop to meet their peers, build and share their skills, and learn about the School of Data way of training people on data skills. What are you waiting for? Read more about School of Data’s Fellowship or Apply now Key Information: Fellowship
  • Available positions: up to 8 fellows, 1 in each of the following countries: Bolivia, Guatemala, Ghana, Indonesia, Kenya, Malawi, Tanzania, The Philippines

  • Application deadline: May 6th, 2018, midnight GMT+0
  • Duration: From May 14th, 2018 to January 31st, 2019
  • Level of activity: 10 days per month
  • Stipend: $1000 USD per month
Key links About diversity and inclusivity School of Data is committed to being inclusive in its recruitment practices. Inclusiveness means excluding no one because of race, age, religion, cultural appearance, sexual orientation, ethnicity or gender. We proactively seek to recruit individuals who differ from one another in these characteristics, in the belief that diversity enriches all that we do. Flattr this!

Apply Now! School of Data’s 2018 Fellowship Programme

- April 16, 2018 in announcement, bolivia, fellowship, ghana, Guatemala, indonesia, kenya, Malawi, philippines, tanzania

UPDATE: the application window has closed. Thanks for applying! School of Data is inviting journalists, data scientists, civil society advocates and anyone interested in advancing data literacy to apply for its 2018 Fellowship Programme, which will run from May 2018 to January 2019. 8 positions are open, 1 in each of the following countries: Bolivia, Guatemala, Ghana, Indonesia, Kenya, Malawi, Tanzania, The Philippines. The application deadline is set on Sunday, May 6th of 2018. If you would like to sponsor a fellowship, please get in touch with School of Data.

The Fellowship

Fellowships are nine-month placements with School of Data for data-literacy practitioners or enthusiasts. During this time, Fellows work alongside School of Data to build an individual programme that will make use of both the collective experience of School of Data’s network to help Fellows gain new skills, and the knowledge that Fellows bring along with them, be it about a topic, a community or specific data literacy challenges. Similarly to previous years, our aim with the Fellowship programme is to increase awareness of data literacy and build communities who together, can use data literacy skills to make the change they want to see in the world. The 2018 Fellowship will continue the work in the thematic approach pioneered by the 2016 class. As a result, we will be prioritising candidates who:
  • possess experience in, and enthusiasm for, a specific area of data literacy training

  • can demonstrate links with an organisation practising in this defined area and/or links with an established network operating in the field
We are looking for engaged individuals who already have in-depth knowledge of a given sector or specific skillsets that can be applied to this year’s focus topics.. This will help Fellows get off to a running start and achieve the most during their time with School of Data: nine months fly by! Read More about the Fellowship Programme

The areas of focus in 2018

We have partnered with Hivos and NRGI to work on the following themes: Procurement and data in the extractives industry (oil, mining, gas). These amazing partner organisations will provide Fellows with guidance, mentorship and expertise in their respective domains.

2018 Fellowship Positions

Bolivia The Fellowship in Bolivia will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: Experience with and interest in community building, experience with the implementation of civic projects with a data or technical component, storytelling skills, and experience with promoting data or technical stories to a wide audience, basic understanding of the public procurement process Guatemala The Fellowship in Guatemala will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: Experience in the planning, coordination and implementation of projects with civil society organisations, the ability to advise and train organisations on working with data and delivering technical projects, basic understanding of the public procurement process Ghana The Fellowship in Ghana with be focused on extractives Data through the Media Development Programme at NRGI. For this position, School of Data is looking for someone with: an interest in supporting or working within the civil society sector, experience working with financial (or related) data for analysis experience as a trainer and/or community builder, interest and/or experience in the extractives sector, demonstrated skills as a data storyteller or journalist Malawi The Fellowship in Malawi will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering technical and data-driven projects, experience with facilitating training activities, experience with data collection projects, basic understanding of the public procurement process Indonesia The Fellowship in Indonesia will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering technical and data-driven projects, experience with facilitating training activities, experience with working with government systems or data. Candidates with the following optional interests and experience will be appreciated: experience with explaining complex topics to varied audiences, experience with user design methodologies, experience with community development The Philippines The Fellowship in The Philippines will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with user-centric research and design methodologies, experience with community-building activities, experience with data storytelling. Candidates with the following optional interests and experience will be appreciated: graphic design skills, experience with delivering trainings Kenya The Fellowship in Kenya will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering data-driven projects, experience with user research and data storytelling, experience with explaining complex topics to varied audiences. Candidates with the following optional interests and experience will be appreciated: interest in or experience with supporting civic projects and civil society organisations, experience with facilitating training activities. Tanzania The Fellowship in Tanzania will be focused on public procurement data through the Open Contracting Programme. For this position, School of Data is looking for someone with: experience with delivering data-driven projects, experience with facilitating training activities, experience with explaining complex topics to varied audiences. Candidates with the following optional interests and experience will be appreciated: experience working with journalists or as a journalist, interest in or experience with supporting civic projects and civil society organisations, experience with writing pedagogical content 9 months to make an impact The programme will run from May to January 2019, and entail up to 10 days a month of time. Fellows will receive a monthly stipend of $1,000 USD a month to cover for their work. What are you waiting for? Read more about School of Data’s Fellowship Key Information: Fellowship
  • Available positions: up to 8 fellows, 1 in each of the following countries: Bolivia, Guatemala, Ghana, Indonesia, Kenya, Malawi, Tanzania, The Philippines

  • Application deadline: May 6th, 2018, midnight GMT+0
  • Duration: From May 14th, 2018 to January 31st, 2019
  • Level of activity: 10 days per month
  • Stipend: $1000 USD per month
Key links About diversity and inclusivity School of Data is committed to being inclusive in its recruitment practices. Inclusiveness means excluding no one because of race, age, religion, cultural appearance, sexual orientation, ethnicity or gender. We proactively seek to recruit individuals who differ from one another in these characteristics, in the belief that diversity enriches all that we do. Flattr this!

Ten Cool Things I Learned at DataJConf

- August 18, 2017 in Events, fellowship

I had a fantastic time at the European Computational and Data Journalism Conference in Dublin on 6-7 July in the company of many like-minded data journalists, academics, and open data practitioners. There were a lot of stimulating ideas shared during the presentations on the first day, the unconference on the second day, and the many casual conversations in between! In this post I’d like to share the ten ideas that stuck with me the most (it was tough to whittle it down to just ten!). Hopefully you’ll find these thoughts interesting, and hopefully they’ll spark some worthwhile discussions about data journalism and storytelling. I’d really love to hear what you have to say about all of this, so please do share any thoughts or observations that you might have below the line! image alt text The European Data and Computational Journalism Conference, Dublin, 6-7 July 2017
  1. ‘Deeper’ data journalism is making a real impact
Marianne Bouchart – manager of the Data Journalism Awards – gave a presentation introducing some of the most exciting award winners of 2017, and talked about some of the most important new trends in data journalism today. Perhaps unsurprisingly, given the electoral rollercoasters of the past year, a lot of great data journalism has been centred around electioneering and other political dramas. Marianne said that “impact” was the theme that ran through the best pieces produced last year, and she really stressed the central role that investigative journalism needs to play in producing strong data-driven stories. She said that impactful investigative journalism is increasingly merging with data journalism, as we saw in projects shedding light on shady anti-transparency moves by Brazilian politicians, investigating the asset-hoarding of Serbian politicians, and exposing irresponsible police handling of sexual assault cases in Canada.
  1. Machine learning could bring a revolution in data journalism
Two academics presented on the latest approaches to computational journalism – journalism that applies machine learning techniques to dig into a story. Marcel Broersma from the University of Groningen presented on an automated analysis of politicians’ use of social media. The algorithm analysed 80,000 tweets from Dutch, British and Belgian politicians to identify patterns of what he called the ‘triangle of political communication’ between politicians, journalists, and citizens. The project wasn’t without its difficulties, though – algorithmically detecting sarcasm still remained a challenge, and the limited demographics of Twitter users meant that this kind of research could only look at how narrow certain segments of society communicated. Jennifer Stark from the University of Maryland looked at the possibilities for algorithms to be biased – specifically looking at Google Image Search’s representations of presidential candidates Hillary Clinton and Donald Trump’s photos during their campaigns. Through the use of an image recognition API that detects emotions, she found that Clinton’s pictures were biased towards showing her appear happier whereas for Trump, both happiness and anger were overrepresented. Although it’s still early days for computational journalism, talks like these hinted at exciting new data journalism methods to come!
  1. There are loads of ways to learn new skills!
The conference was held at the beautiful University College Dublin, where a brand new master’s program in data journalism is being launched this year. We also heard from one of the conference organisers, Martin Chorley, about Cardiff University’s Master’s in Computational and Data Journalism, which has been going strong for three years, and has had a great track record of placing students into employment. But formal education isn’t the only way to get those cutting edge data journo skills! One of the conference organisers also presented the results of a worldwide survey of data journalists, taking in responses from 180 data journalists across 44 countries. One of the study’ most notable findings was that only half of respondents had formal training in data journalism – the rest picked up the necessary skills all by themselves. Also, when asked how they wanted to further their skills, more respondents said they wanted to brush up on their skills in short courses rather than going back to school full-time.
  1. Want good government data? Be smart (and be charming)!
One of the most fascinating parts of the conference for me was learning about the different ways data journalists obtained data for their projects. Kathryn Tourney from The Detail in Northern Ireland found Freedom of Information requests useful, but with the caveat that you really needed to know the precise structure of the data you are requesting in order to get the best data. Kathryn would conduct prior research on the exact schemas of government databases and work to get hold of the forms that the government used to collect the data she wanted before making the actual FOI requests. This ensured that there was no ambiguity about what she’d receive on the other side! Conor Ryan from Ireland’s RTÉ found that he didn’t need to make FOI requests to do deep investigative work, because there was already a lot of government data “available” to the public. The catch was that this data was often buried behind paywalls and multiple layers of bureaucracy. Conor stressed the importance of ensuring that any data sources RTÉ managed to wrangle were also made available in a more accessible way for future users. One example related to accessing building registry data in Ireland, where originally a €5 fee existed for every request made. Conor and his team pointed out this obstacle to the authorities and persuaded them to change the rules so that the data would be available in bulk in the future. Lastly, during the unconference one story from Bulgaria really resonated with my own experiences trying to get a hold of data from governments in closed societies. A group of techies offered the Bulgarian government help with an array of technical issues, and by building relationships with staff on the ground – as well as getting the buy-in of political decision makers – they were able to get their hands on a great deal of data that would have forever remained inaccessible if they’d gone through the ‘standard’ channels for accessing public information.
  1. The ethics of data sharing are tricky
The best moments at these conferences are the ones that make you go: “Hmm… I never thought about it that way before!”. During Conor Ryan’s presentation, he really emphasized the need for data journalists to consider the ethics of sharing the data that they have gathered or analysed. He pointed out that there’s a big difference between analysing data internally and reporting on a selected set of verifiable results, and publishing the entire dataset from your analysis publicly. In the latter case, every single row of data becomes a potential defamation suit waiting to happen. This is especially true when the dataset involved is disaggregated down the level of individuals!
  1. Collaboration is everything
Being a open data practitioner means that my dream scenarios are collaborations on data-driven projects between techies, journalists and civil society groups. So it was really inspiring to hear Megan Lucero talk about how The Bureau Local (at the Bureau of Investigative Journalism) has built up a community of civic techies, local journalists, and civil society groups across the UK. Even though The Bureau Local was only set up a few months ago, they quickly galvanized this community around the 2017 UK general elections, and launched four different collaborative investigative data journalism projects. One example is their piece on targeted ads during the election campaign, where they collaborated with the civic tech group Who Targets Me to collect and analyse data about the kinds of political ads targeting social media users. I’d love to see more experiments like The Bureau Local emerging in other countries as well! In fact, one of the main purposes of Open and Shut is precisely to build this kind of community for folks in closed societies who want to collaborate on data-driven investigations. So please get involved! image alt text Who Targets Me? Is an initiative working to collect and analyse data about the kinds of political ads targeting social media users.
  1. Data journalism needs cash – so where can we find it?
It goes without saying these days that journalism is having a bad time of it at the moment. Advertising and subscription revenues don’t pull in nearly as much cash as the used to. Given that pioneering data-driven investigative journalism takes a lot of time and effort, the question that naturally arises is: “where do we get the money for all this?”. Perhaps unsurprisingly, no-one at DataJConf had any straightforward answers to this question. A lot of casual conversations in between sessions drifted onto the topic of funding for data journalism, and lots of people seemed worried that innovative work in the field is currently too dependent on funding from foundations. That being said, attendees also shared stories about interesting funding experiments being undertaken around the world, with the Korean Center for Investigative Journalism’s crowdfunding approach gaining some interest.
  1. Has data journalism been failing us?
In the era of “fake news” and “alternative facts”, a recurring topic in many conversations was about whether data journalism actually had any serious positive impacts. During the unconference discussions, some of us ended up being sucked into the black hole question of “What constitutes proper journalism anyway?”. It wasn’t all despair and navel-gazing, however, and we definitely identified a few concrete things that could be improved. One related to the need to better represent uncertainty in data journalism. This ties into questions of improving the public’s data literacy, but also of traditional journalism’s tendency to present attention-grabbing leads and conclusions without doing enough to convey complexity and nuance. People kept referencing FiveThirtyEight’s election prediction page, which contained a sophisticated representation of the uncertainty in their modelling, but hid it all below the fold – an editorial decision, it was argued, that lulled readers into thinking that the big number that they saw at the top of the page was the only thing that mattered. image alt text FiveThirtyEight’s forecast of the 2016 US elections showed a lot of details below the fold about their forecasting model’s uncertainty, but most readers just looked at the big percentages at the top. Another challenge identified by attendees was that an enormous amount of resources were being deployed to preach to the choir instead of reaching out to a broader base of readers. The unconference participants pointed out that a lot of the sophisticated data journalism stories written in the run-up to the 2016 US elections were geared towards partisan audiences. We agreed that we needed to see more accessible, impactful data stories that were not so mired in party politics, such as ProPublica’s insightful piece on rising US maternal mortality rates.
  1. Data journalism can be incredibly powerful in the Global South
Many of the talks were about data journalism as it was practised in Western countries – with one notable exception. Eva Constantaras, who trains investigative data journalism teams in the Global South, held a wonderful presentation about the impactfulness of data journalism in the developing world. She gave the examples of IndiaSpend in India and The Nation in Kenya, and spoke about how their data-driven stories worked to identify problems that resonated with the public, and explain them in an accessible and impactful way. Election coverage in these two examples shared by Eva focused on investigating the consequences of the policy proposals of politicians, engaging in fact-checking, and identifying the kinds of problems that were faced by voters in reality. Without the burden of partisan echo-chambers, and because data journalism is still very new and novel in many parts of the world, data journalism could end up having a huge impact on public debate and storytelling in the Global South. Watch this space! image alt text Kenya’s *The Nation has been producing data-driven stories more and more frequently, such as **this piec*e* on Kenya’s Eleventh Elections in August 2017*
  1. Storytelling has to connect on a human level
If there was one recurring theme that I heard throughout the conference about what makes data journalism impactful, it was that the data-driven story has to connect on a human level. Eva had a slide in her talk with a quote from John Steinbeck about what makes a good story: “If a story is not about the hearer he [or she] will not listen… A great lasting story is about everyone, or it will not last. The strange and foreign is not interesting – only the deeply personal and familiar.” Embed youtube video: https://youtu.be/lFEHt_gOcwo “I want loads of money” — Councillor Hugh McElvaney caught on hidden camera video from RTÉ Conor from RTÉ also drove the same point home. After his team’s extensive data-driven investigative work revealed corruption in Irish politics, the actual story that they broke involved a hidden-camera video of an undercover interview with one of these politicians. This video highlighted just one datapoint in a very visceral way, which ultimately resonated more with the audience than any kind of data visualisation could.
I could go on for longer, but that’s probably quite enough for one blog post! Thanks for reading this far, and I hope you managed to gain some nice insights from my experiences at DataJConf. It was a fascinating couple of days, and I’m looking forward to building upon all of these exciting new ideas in the months ahead! If any of these thoughts have got you excited, curious (or maybe even furious) we’d love to hear from you below the line. *Open & Shut is a project from the **Small Medi*a* team. Small Media are an organisation working to support freedom of information in closed societies, and are behind the portal **Iran Open Dat*a*.* Flattr this!

Ten Cool Things I Learned at DataJConf

- August 18, 2017 in Events, fellowship

This article was cross-posted from its original location at the Open and Shut blog I had a fantastic time at the European Computational and Data Journalism Conference in Dublin on 6-7 July in the company of many like-minded data journalists, academics, and open data practitioners. There were a lot of stimulating ideas shared during the presentations on the first day, the unconference on the second day, and the many casual conversations in between! In this post I’d like to share the ten ideas that stuck with me the most (it was tough to whittle it down to just ten!). Hopefully you’ll find these thoughts interesting, and hopefully they’ll spark some worthwhile discussions about data journalism and storytelling. I’d really love to hear what you have to say about all of this, so please do share any thoughts or observations that you might have below the line! image alt text The European Data and Computational Journalism Conference, Dublin, 6-7 July 2017
  1. ‘Deeper’ data journalism is making a real impact
Marianne Bouchart – manager of the Data Journalism Awards – gave a presentation introducing some of the most exciting award winners of 2017, and talked about some of the most important new trends in data journalism today. Perhaps unsurprisingly, given the electoral rollercoasters of the past year, a lot of great data journalism has been centred around electioneering and other political dramas. Marianne said that “impact” was the theme that ran through the best pieces produced last year, and she really stressed the central role that investigative journalism needs to play in producing strong data-driven stories. She said that impactful investigative journalism is increasingly merging with data journalism, as we saw in projects shedding light on shady anti-transparency moves by Brazilian politicians, investigating the asset-hoarding of Serbian politicians, and exposing irresponsible police handling of sexual assault cases in Canada.
  1. Machine learning could bring a revolution in data journalism
Two academics presented on the latest approaches to computational journalism – journalism that applies machine learning techniques to dig into a story. Marcel Broersma from the University of Groningen presented on an automated analysis of politicians’ use of social media. The algorithm analysed 80,000 tweets from Dutch, British and Belgian politicians to identify patterns of what he called the ‘triangle of political communication’ between politicians, journalists, and citizens. The project wasn’t without its difficulties, though – algorithmically detecting sarcasm still remained a challenge, and the limited demographics of Twitter users meant that this kind of research could only look at how narrow certain segments of society communicated. Jennifer Stark from the University of Maryland looked at the possibilities for algorithms to be biased – specifically looking at Google Image Search’s representations of presidential candidates Hillary Clinton and Donald Trump’s photos during their campaigns. Through the use of an image recognition API that detects emotions, she found that Clinton’s pictures were biased towards showing her appear happier whereas for Trump, both happiness and anger were overrepresented. Although it’s still early days for computational journalism, talks like these hinted at exciting new data journalism methods to come!
  1. There are loads of ways to learn new skills!
The conference was held at the beautiful University College Dublin, where a brand new master’s program in data journalism is being launched this year. We also heard from one of the conference organisers, Martin Chorley, about Cardiff University’s Master’s in Computational and Data Journalism, which has been going strong for three years, and has had a great track record of placing students into employment. But formal education isn’t the only way to get those cutting edge data journo skills! One of the conference organisers also presented the results of a worldwide survey of data journalists, taking in responses from 180 data journalists across 44 countries. One of the study’ most notable findings was that only half of respondents had formal training in data journalism – the rest picked up the necessary skills all by themselves. Also, when asked how they wanted to further their skills, more respondents said they wanted to brush up on their skills in short courses rather than going back to school full-time.
  1. Want good government data? Be smart (and be charming)!
One of the most fascinating parts of the conference for me was learning about the different ways data journalists obtained data for their projects. Kathryn Tourney from The Detail in Northern Ireland found Freedom of Information requests useful, but with the caveat that you really needed to know the precise structure of the data you are requesting in order to get the best data. Kathryn would conduct prior research on the exact schemas of government databases and work to get hold of the forms that the government used to collect the data she wanted before making the actual FOI requests. This ensured that there was no ambiguity about what she’d receive on the other side! Conor Ryan from Ireland’s RTÉ found that he didn’t need to make FOI requests to do deep investigative work, because there was already a lot of government data “available” to the public. The catch was that this data was often buried behind paywalls and multiple layers of bureaucracy. Conor stressed the importance of ensuring that any data sources RTÉ managed to wrangle were also made available in a more accessible way for future users. One example related to accessing building registry data in Ireland, where originally a €5 fee existed for every request made. Conor and his team pointed out this obstacle to the authorities and persuaded them to change the rules so that the data would be available in bulk in the future. Lastly, during the unconference one story from Bulgaria really resonated with my own experiences trying to get a hold of data from governments in closed societies. A group of techies offered the Bulgarian government help with an array of technical issues, and by building relationships with staff on the ground – as well as getting the buy-in of political decision makers – they were able to get their hands on a great deal of data that would have forever remained inaccessible if they’d gone through the ‘standard’ channels for accessing public information.
  1. The ethics of data sharing are tricky
The best moments at these conferences are the ones that make you go: “Hmm… I never thought about it that way before!”. During Conor Ryan’s presentation, he really emphasized the need for data journalists to consider the ethics of sharing the data that they have gathered or analysed. He pointed out that there’s a big difference between analysing data internally and reporting on a selected set of verifiable results, and publishing the entire dataset from your analysis publicly. In the latter case, every single row of data becomes a potential defamation suit waiting to happen. This is especially true when the dataset involved is disaggregated down the level of individuals!
  1. Collaboration is everything
Being a open data practitioner means that my dream scenarios are collaborations on data-driven projects between techies, journalists and civil society groups. So it was really inspiring to hear Megan Lucero talk about how The Bureau Local (at the Bureau of Investigative Journalism) has built up a community of civic techies, local journalists, and civil society groups across the UK. Even though The Bureau Local was only set up a few months ago, they quickly galvanized this community around the 2017 UK general elections, and launched four different collaborative investigative data journalism projects. One example is their piece on targeted ads during the election campaign, where they collaborated with the civic tech group Who Targets Me to collect and analyse data about the kinds of political ads targeting social media users. I’d love to see more experiments like The Bureau Local emerging in other countries as well! In fact, one of the main purposes of Open and Shut is precisely to build this kind of community for folks in closed societies who want to collaborate on data-driven investigations. So please get involved! image alt text Who Targets Me? Is an initiative working to collect and analyse data about the kinds of political ads targeting social media users.
  1. Data journalism needs cash – so where can we find it?
It goes without saying these days that journalism is having a bad time of it at the moment. Advertising and subscription revenues don’t pull in nearly as much cash as the used to. Given that pioneering data-driven investigative journalism takes a lot of time and effort, the question that naturally arises is: “where do we get the money for all this?”. Perhaps unsurprisingly, no-one at DataJConf had any straightforward answers to this question. A lot of casual conversations in between sessions drifted onto the topic of funding for data journalism, and lots of people seemed worried that innovative work in the field is currently too dependent on funding from foundations. That being said, attendees also shared stories about interesting funding experiments being undertaken around the world, with the Korean Center for Investigative Journalism’s crowdfunding approach gaining some interest.
  1. Has data journalism been failing us?
In the era of “fake news” and “alternative facts”, a recurring topic in many conversations was about whether data journalism actually had any serious positive impacts. During the unconference discussions, some of us ended up being sucked into the black hole question of “What constitutes proper journalism anyway?”. It wasn’t all despair and navel-gazing, however, and we definitely identified a few concrete things that could be improved. One related to the need to better represent uncertainty in data journalism. This ties into questions of improving the public’s data literacy, but also of traditional journalism’s tendency to present attention-grabbing leads and conclusions without doing enough to convey complexity and nuance. People kept referencing FiveThirtyEight’s election prediction page, which contained a sophisticated representation of the uncertainty in their modelling, but hid it all below the fold – an editorial decision, it was argued, that lulled readers into thinking that the big number that they saw at the top of the page was the only thing that mattered. image alt text FiveThirtyEight’s forecast of the 2016 US elections showed a lot of details below the fold about their forecasting model’s uncertainty, but most readers just looked at the big percentages at the top. Another challenge identified by attendees was that an enormous amount of resources were being deployed to preach to the choir instead of reaching out to a broader base of readers. The unconference participants pointed out that a lot of the sophisticated data journalism stories written in the run-up to the 2016 US elections were geared towards partisan audiences. We agreed that we needed to see more accessible, impactful data stories that were not so mired in party politics, such as ProPublica’s insightful piece on rising US maternal mortality rates.
  1. Data journalism can be incredibly powerful in the Global South
Many of the talks were about data journalism as it was practised in Western countries – with one notable exception. Eva Constantaras, who trains investigative data journalism teams in the Global South, held a wonderful presentation about the impactfulness of data journalism in the developing world. She gave the examples of IndiaSpend in India and The Nation in Kenya, and spoke about how their data-driven stories worked to identify problems that resonated with the public, and explain them in an accessible and impactful way. Election coverage in these two examples shared by Eva focused on investigating the consequences of the policy proposals of politicians, engaging in fact-checking, and identifying the kinds of problems that were faced by voters in reality. Without the burden of partisan echo-chambers, and because data journalism is still very new and novel in many parts of the world, data journalism could end up having a huge impact on public debate and storytelling in the Global South. Watch this space! image alt text Kenya’s The Nation has been producing data-driven stories more and more frequently, such as this piece on Kenya’s Eleventh Elections in August 2017*
  1. Storytelling has to connect on a human level
If there was one recurring theme that I heard throughout the conference about what makes data journalism impactful, it was that the data-driven story has to connect on a human level. Eva had a slide in her talk with a quote from John Steinbeck about what makes a good story: “If a story is not about the hearer he [or she] will not listen… A great lasting story is about everyone, or it will not last. The strange and foreign is not interesting – only the deeply personal and familiar.” “I want loads of money” — Councillor Hugh McElvaney caught on hidden camera video from RTÉ Conor from RTÉ also drove the same point home. After his team’s extensive data-driven investigative work revealed corruption in Irish politics, the actual story that they broke involved a hidden-camera video of an undercover interview with one of these politicians. This video highlighted just one datapoint in a very visceral way, which ultimately resonated more with the audience than any kind of data visualisation could.
I could go on for longer, but that’s probably quite enough for one blog post! Thanks for reading this far, and I hope you managed to gain some nice insights from my experiences at DataJConf. It was a fascinating couple of days, and I’m looking forward to building upon all of these exciting new ideas in the months ahead! If any of these thoughts have got you excited, curious (or maybe even furious) we’d love to hear from you below the line. Open & Shut is a project from the Small Media team. Small Media are an organisation working to support freedom of information in closed societies, and are behind the portal Iran Open Dat*a. Flattr this!