2019 Argentine Election: the challenge of reinventing

Category: Best visualization (small and large newsrooms)

Country/area: Argentina

Organisation: LA NACION

Organisation size: Big

Publication date: 8 Jan 2019

Credit: Pablo Loscri,Florencia Fernández Blanco,Nicolás Bases,Cristian Bertelegni,Gastón de la Llana,Mariana Trigo Viera,Nicolás Rivera,Gise Ferro,Alejandra Bliffeld,Carlos Araujo,Florencia Abd,Juana Copello,Florencia F. Altube,Bianca Pallaro,Delfina Arambillet

Project description:

How to innovate for an event that takes place every four years? How can we distinguish ourselves from the competition and from our past selves? The key was understanding the needs of the audience and generating DATA visualizations through innovative techniques surrounding data acquisition and analysis. From interactive maps and electoral calculators to specific applications and pieces targeting young audiences that conveyed key journalistic findings in a clear and concise way. 


Impact reached:

It is important to always keep the journalism thermometer intact and detect the new needs that arise in the moment. Former president Mauricio Macri was unexpectedly trounced in primary elections. So everybody was asking: would he have the chance to reach a second round and remain in power, or would Alberto Fernández win in the first round? To answer the question, we created a custom calculator that became so popular that according to trusted soruces, even ministers and advisors from the Government House used the application to make arrangements about the future of the campaign. As regards to metrics, the application was highly valued by the audience with almost 500,000 page views and the electoral maps with the results and all its features hit the record of 685,173 page views. Maps and visualizations about the new composition of the Congress and the general presidential results were used by several TV argentine channels with a national range. And the visualisations that targetted young audiences: Google search interest, Questions selected by our social media account followers and answered by all 9 candidates, the debate game among others, became the most shared articles through social media. 



Techniques/technologies used:

The applications created for the election coverage were divided into two groups: the ones updated in real time and the static. For the election maps and the calculation of seats in the Congress (both applications received data from the election counting in real time) a CMS was developed in Python with Diango library. Data of candidates and parties that were going to participate in the elections were previously loaded though CMS. Data was stored in a Postgres database and, during the election, in real time, data was crossed with the information of the official results of the elections and used by the different applications created. Data parsing was made in the back-end with Python. The front-end of all the election web apps was made with Javascript frameworks. Vue.js was mainly used for real time apps, for DOM and for data binding. The Javascript library called D3js was also used to create maps and graphics. The rest of the static applications were made in HTML and Javascript, Vue.js was also used for the DOM. In some other applications it was used jquery, and leaflet.js was used to create tile maps.

What was the hardest part of this project?

The most difficult challange of the project arose when we found oyt that the company in charge of the election counting in Argentina was changed. In order to automate the results and display them in real time, the team needs to have data structure with a proper advance to be able to adapt data to our application software. The night before the elections we still didn’t had that information and, in many cases, we had to inform them about errors or failures in the information sent. Expecting, perhaps, some complications with data automation during the live election night, especially in the first hours when all the media organizations compete for the scoop, we decided to implement an information system to feed our result boards by hand to avoid depending on the official submission that showed deficiencies. To achieve an immediate publication, a group of 10 persons were in charge of the data uploading and we were the first to reflect the results since the automation did not work and the other media did not have this alternative plan. When the sending was stabilized, we “plugged in” with the company and the rest of the counting was automatic, which allowed us to feed not only the result boards but all the pieces that involved large volumes of data such as the maps of the entire country, the new composition of the Congress in real time and the comparative “weighing scales” of the two main political forces with respect to the previous elections.

What can others learn from this project?

There are four lessons or, at least, suggestions that we could offer based on our experience with this project. On the one hand, to sharpen perception of reality to make data driven reporting according to the needs of the audience. Secondly, understand that we can be facilitators/means and bring solutions or provide answers through technologies that serve the needs of those who consume us or of those who may start to do so if we show them something new, that has quality and is useful. Thirdly, and based on the story about the disadvantages of automating the results data, it is important to always have a Plan B, especially when the success or failure of an application does not depend on us but on a third party (in this case the company in charge of the election counting). Finally, it is worth mentioning the crossover between platforms such as social media. It is really important to focus on the future readers needs, especially young audiences. 

Project links: