Visualizing the local – efficient data journalism for small newsrooms

2020 Shortlist

Visualizing the local – efficient data journalism for small newsrooms

Category: Innovation (small and large newsrooms)

Country/area: Sweden

Organisation: Approximately 170 different media including Borås Tidning, Lokaltidningen, Västerås Tidning, Aftonbladet, Blekinge läns tidning, Sveriges radio

Organisation size: Small

Publication date: 9 Feb 2019

Credit: Anna Norberg, Gunnar Jacobsson, Fredrik Johansson, Jesper Eriksson, Robin Orling, Marie Ericsson, Per Larsson, Jack Carlsson, Ulf Fahlén

Project description:

Nyhetsbyrån Siren uses open data, FOI requests and public documents to create data journalism concerning the most pressing issues in Swedish society. On a daily basis we collect, clean and share datasets with over 170 newsrooms all over the country, so that their journalists can write stories about issues affecting their local communities. In addition to the dataset, we write out the key findings along with an expert interview to provide extra context. For every dataset we create a series of locally customized graphics for publication. This is done using limited editorial resources, most often only one person.

Impact reached:

The project helps strengthen local and national news rooms, in times of lower budgets and staff reductions. Journalists all over the country can offer their readers/listeners/viewers several data driven stories each week with customized graphics. And they do. Siren’s stories are picked up and published in many different media outlets on a daily basis. It is not uncommon that our news stories dominate the national media as well. During autumn 2019, Siren has put a light on topics such as the falling numbers of physicians with a specialist degree, different opportunities for children to learn music and culture depending on where they live, the regional impact on a much discussed tax reduction for those with the highest income, the most polluting industries throughout the country and the fact that reports of domestic violence are increasing during and after the Christmas holidays. To name just a few.

Techniques/technologies used:

Excel and R are used for collecting and sorting data gathered from open sources and public offices. The project has been ongoing for three years and up until september 2019 we used Datawrapper and Infogram for creating graphics. In september we started using R scripts instead for visualizing the data and creating locally customized graphics. By using R scripts we have much more control over the appearance, it is much easier to update data or text in all the graphics at the same time or to download and name all graphics in batches. Our own editorial system Sirenen, launched in spring 2019, is used for the distribution of all our news to other media.

What was the hardest part of this project?

We are a small news agency with high journalistic ambitions, especially when it comes to producing quality stories for struggling local media. But how to do that, reaching the whole of our sparsely populated country, with just nine editorial staff? The answer is in part data journalism. Using reproducable R scripts to easily multiply and customize graphics saves us a lot of time and effort. Our data journalist Anna Norberg worked out how we could fit R into our editorial process and is now spreading the knowledge to the rest of the team, so that everyone at Siren easily can contribute with their own data journalism stories.

How Sweden is divided geographically, with 21 counties and 290 municipalities, is another problem. The municipalities are not evenly distributed, for example over 45 municipalities belong to Västra Götaland county, but Gotland county has only one municipality. We usually make one chart for each county, showing data for all the municipalities in that county. This works fine for most of the counties, but for others the charts have not been very useful, either showing only one value or too many. With Datawrapper it has never been an option to trying to make individual charts for each municipality. That would be too time consuming. But with R we can now start to address this problem and just as easily produce 290 charts as 21, and in no time.

What can others learn from this project?

If you dare try new methods and open source tools, and actually invest the time it takes to master them properly, you can save time and accomplish amazing things. Nyhetsbyrån Siren have limited resources and can not involve outside developers in everything we do. By letting our datajournalist Anna Norberg spend some time to develop R skills, we can now visualize our data journalism stories more efficiently than we could ever have imagined a year ago.