2022 Winner


Country/area: Argentina

Organisation: LA NACION (Argentina)

Organisation size: Big

Jury’s comments:

Once again in 2021, La Nacion Data served as a model for news organizations around the world. In a country without a freedom of information law, La Nation Data continues to find creative ways to acquire data and tell stories of significance. Whether through crowdsourcing or collaboration, La Nacion simply finds new and creative ways to get the story. The jury singled out La Nacion’s creative use of computer vision, AI and crowdsourcing to monitor paper election ballots in a country where votes are counted manually. They found no fraud but did put politicians on notice that this ground journalists, students and NGOs were watching. Truly incredible work.

Cover letter:

In October 2010, after an ONA pre conference workshop “Data Journalism for Beginners”, we launched LA NACION DATA as a newsroom strategy. We dreamt big, started small, and never stopped.

Our beginnings were really challenging. Argentina was a country without FOIA law and no open data policies.

We started gathering allies from within the company but also making bonds with the incipient Hacks/Hackers, open data, open government and data science initiatives. To accelerate change we hosted 4 Datafest public events in 2012, 2013, 2014 and 2016 and strengthened  learnings on opening, analyzing and visualizing data of public interest generated by different stakeholders (government, multilateral organizations, academia, media, NGos, etc).

Collaboration is in our DNA. We started in 2012 for our news application on statement of assets working with transparency NGOs until the launch of our Environment Legislator Monitor in 2021. We also partner with other  newsrooms in national (Chequeado), latinamerican (Red Palta) and international (ICIJ, Journalism AI collab) areas.

We went massive with crowdsourcing developing our open source, online and collaborative platform Vozdata with the support of Knight Mozilla OpenNews and Civicus. Since 2014, we have used it to digitize and make databases from scratch with 10.000 PDFs of Senate expenses, 40.000 phone interceptions’ audios and 16.000 Elections documents PDFs.  This year we used the platform to audit a computer vision algorithm preselected 30.000 PDFs documents with the help of 300 volunteers.

Our scope includes data driven investigations, data as a service and “data-teinment”.

We are a team of 12  journalists,  data scientists, developers and designers. We  usually work in  collaboration with the Graphics department for big projects.

Our main inspiration through the last 11 years came from  Propublica, The Guardian UK, Los Angeles Times, NYTimes, WaPo, Civio (Spain), The Economist, Reuters, Ojo Publico, Texty.

We usually got ideas from conferences like NICAR, ONA, GEN Summit, Abrelatam, Hacks/Hackers Argentina, WIDS and data journalism books.

Mentors: Knight Mozilla OpenNews, EJC, JSK at Stanford University, Prof. Rosental Alves and lately from, PolisLSE Journalism AI Collab.

Description of portfolio:

2021 Highlights

[Climate & Biodiversity Crisis]: Our herculean effort to visualize impossible unstructured data from Congress bills projects related to climate and biodiversity crisis. The Legislative Environmental Monitor is based on a huge database made from scratch and updated periodically from unstructured text. This tool is helpful to understand the law enactment process, monitor the advance of bills in an opaque context and provide a public service for the audience to demand specific measures against the climate & biodiversity crisis. This application was made by LA NACION Data together with four NGOs ‒one of them, a specialist in legislative matters‒ and the other three are local references in environmental topics.

[AI + Crowdsourcing] Our first Computer Vision project for monitoring Elections documentation in a country where votes are counted manually and there are multiple paper options as political parties participate in primaries or general elections, (it could go to + 15 different options to count in some districts). The purpose of the project was to monitor telegrams with computer vision and then validate data provided by the algorithm using crowdsourcing.

[AI + Text]: Our first NLP and machine learning project analyzing 600 hundred lyrics from trap music. Trap is one of the most popular music genres among young people in Argentina. We used Natural Language Processing (NLP) tools to analyze their lyrics, demystify the concepts around the genre and understand what these artists, who bring about such a furor among the new generations, are talking about.

[Legislative Elections pack]. Electoral coverage presents the challenge of innovating on an event that takes place in the country every two years. At LA NACION we developed more than ten pieces with great visual and interactive power to provide a public service and guarantee a differential consumption experience for the newspaper’s audience. At the same time, we guarantee the detailed analysis of the large volume of data (more than 24.4 million votes distributed in 379 departments, 135 districts and 15 municipalities) and its comparison with previous elections, whose data are stored in a large historical database.

[Corruption in Vaccination against Covid]: The vaccination campaign against COVID in Argentina had irregularities since its inception. Less than two months after it began, a corruption scandal known as the “VIP vaccination center” came to light: government officials, family members and those closest to the government had received doses before the prioritized groups. We decided to follow the campaign and the vaccination system from two different sides. On the one hand, control the number of doses received, distributed and applied. On the other hand, monitor the implementation of public policies by combining public data and many requests for access to public information as main sources.

[Real Economy Monitor (MER)]: Argentina is within the Top 5 countries in the world regarding inflation. Our domestic and hyperregulated economy lives with 7 different types of US dollars conversion rate to Argentina’s pesos. MER is an invaluable tool for our readers to follow for free more than 80 indicators. Data as a service maintained by LN Data team. And we got sponsored by a major commercial bank!

Notice: We are using the links boxes provided by SIGMA Awards to put info in official links in spanish to experience the original reporting and another link for the project with behind the scenes explanation in english, as needed.

Project links: