Sonar emerged out of the “X-raying the politics” project, supported by the Google Digital News Initiative fund. The project came from an idea to bring political topics to users and show them how it can affect their lives. It was also based on our conviction that tech tools and knowledge, combined with extraordinary journalism and open source databases, can change public discourse and politics for the better.

Sonar was developed through a collaboration between Laboratorium EE and Polish daily newspaper Gazeta Wyborcza, published by Agora Media Group. Not just another regular politics Web site, Sonar is a bold experiment that breaks the mold of what is usually presented by media companies.

Sonar is mostly a data-driven digital publication, but it is also a tool that supports Gazeta Wyborcza’s political journalism and empowers users to keep tabs over time on the specific issues, candidates, and politicians that affect their lives and communities.

At a glance, readers can quickly learn about a politician's positions on issues and voting track record.
At a glance, readers can quickly learn about a politician's positions on issues and voting track record.

At first glance, Sonar looks more like a music site such as Spotify. Profiles of politicians resemble profiles of artists. But instead of album release dates, users can view information such as:

  • When each of the 460 members of parliament, 100 senators, and other politicians were elected.
  • What are their roles within their political parties.
  • How they have voted on major decisions (whether they were for, against, or withdrew from voting).

Machine learning is one of Sonar’s most important aspects because it compares people’s preferences and is able to identify “matches” with politicians and other users. While browsing through politicians’ profiles, the system can identify those that are closest in opinion or those that are mostly opposite in terms of political views.

Sonar checks, counts, analyses, shows, and informs — but never comments. Data for Sonar is collected by bots designed specifically for visiting various Web sites (i.e. the parliament Web site) and programming interfaces (APIs). The tool has an advanced, built-in system to control the bots as they browse the Web for information about politicians.

Similar to Spotify for music artists or Facebook for friends and family, Sonar helps users find politicians and other users that share their views.
Similar to Spotify for music artists or Facebook for friends and family, Sonar helps users find politicians and other users that share their views.

Thanks to unique access to the Google Trends API, Sonar lets users keep track of their interest in politics on the Web. Even though users could check politicians in Google Trends manually, Sonar does it automatically every 15 minutes by checking more than 600 names.

The Google data is provided on a relative scale, depending on the query structure. In practice, it makes it quite difficult to compare the popularity of politicians — or even impossible when a sudden popularity surge pops it out of scale. Sonar solves this problem by adding a reference word to the queries about the popularity of politicians, thereby helping set the boundaries of the scale.

Sonar components communicate only through the API and as a result it is more flexible and easier to modify. Unusual situations, such as change of politician’s surname, also were taken into account. The Web site links the correct data, even when system administrators change the surname of the deputy or when data on the parliament Web site is still entered under the old name. Similarly, there is no problem collecting information when a politician has a change of party affiliation. The system can also export data (both CSV/XLS files, as well as programming interfaces) and work with them conveniently (convert, search, etc.)

Data is collected from a variety of sources and automatically displayed in a user-friendly format.
Data is collected from a variety of sources and automatically displayed in a user-friendly format.

Sonar has been a fully operating information source for Gazeta Wyborcza subscribers and as as an added value to its digital subscribers. Its content is not free, but comes in a bundle and can be accessed only after subscribing to Agora’s flagship product.

It not only supports production of articles, but also brings entire blocks of content and functions to the Web site, such as politician profiles. Specific topics or politicians can be followed on Sonar, just like friends or celebrities on Facebook or Twitter.

Users also have the choice to customise the site to their own interests by selecting topics that are important to them and receiving notifications. Users have full control over which data the site provides to them. Their regular answers to questions in quizzes lets Sonar build a “political mirror” that uses algorithms to evaluate which political party users are close to — or they can use this to discover politicians and other users with similar views. Sonar is also a fact-checking tool with articles prepared by journalists from Gazeta Wyborcza.

The Sonar project was conducted through a close partnership with the Agora Wyborcza.pl team, which means Laboratorium EE has been present at its every single stage, from the birth of the idea, through design, UX, coding, data analysis, and the development of business and marketing strategy.

On each of those levels, experts from Laboratorium EE worked together with the Gazeta Wyborcza team on designing the Web site and the project involved a team of programmers from wyborcza.pl. Thanks to this collaboration, a modern service for users was created, but it also was fully integrated with Agora’s IT and marketing systems as well.

One of project’s biggest advantages is its ease of scale. The knowledge gained through the process combined with Laboratorium EE’s experience with data make this project relatively easy to be implemented in additional markets moving forward.