Project description

RTL Nieuws is one of the largest news organizations in The Netherlands. We produce daily news broadcasts across two tv channels, as well as online journalism across several websites and social media platforms.

Within the investigative team at RTL Nieuws, two data journalists work full time to produce hard-hitting news stories using data and DDJ techniques, as well as innovative, interactive tools, maps and graphics.

We have a broad audience, with viewers and readers throughout the country, with different ages, education levels and family situations. Because of this, we focus on subjects that are of interest to large parts of the Dutch population. Prime subject areas include healthcare, safety, education and personal finance.

The stories we submitted in this portfolio:

– School quality: We analyzed standardized testing scores for every primary school in The Netherlands (over 6.000), correcting for the different backgrounds of the student population. This gives a good insight into the quality of education at each school. Our method has piqued the interest of the Dutch Ministry of Education.

– Environmental noise: We correlated an environmental noise map by the Dutch public health service to a dataset of every residential address and compared the resulting noise levels with new WHO guidelines. This way we found out that noise levels are at an unhealthy level for more than a third of all homes.

– Burglaries near motorways: analysis of open police data showed that burglaries are much more prevalent near motorways. Police confirmed this theory.

– Social housing: scraping the websites of social housing agencies, we were able to show that waitlists for social housing are extremely long. The average new tenant in the past year had been registered with the social housing agency for almost nine years. In some cities, the average was over twenty years.

– Provincial roads: We analyzed traffic accidents on provincial (secondary) roads in The Netherlands and gave all of them a star rating based on their safety (using a home-made model). Our research showed that these roads are the most dangerous in the country. They comprise only six percent of roads in The Netherlands, yet twenty percent of traffic deaths occur on these roads. Road maintenance was severely underfunded.

– Traffic accidents near schools: Using traffic accident data and a questionnaire sent to Dutch schools, we found out that nine out of ten schools experience traffic safety issues. In the past three years, more than 10.000 accidents happened in the immediate vicinity of a school.

– Healthcare cost differences: we requested health insurance data and found out that there are large differences in healthcare costs for children across neighborhoods. Overall, we found that children who grow up in poorer areas are less healthy. These differences start at birth: healthcare costs for babies can differ by as much as 1000 euros per child per year between neighborhoods.

What makes this project innovative?

The fact that we work for a broadcast news program, means we have to translate complicated data investigations to compelling tv news items, that both contain a strong news angle and appeal to our viewers. To do this, the data team at RTL Nieuws produces thorough, data-driven investigative reporting, and also translates data in a way that's clear and easy to understand for our viewers and readers. We pride ourselves in creating content that's useful for our target audience. The question we always want to be able to answer: 'what does this mean for me?'. To that end, we create compelling online interactives, such as detailed maps, preferably on the level of neighborhoods or even individual homes or schools. We always try to enable readers to look up their specific situation. For instance, this tax season we created tools that allowed readers to find out exactly how much tax they pay. We also differentiate ourselves by working together with third parties, both as data sources and to improve our methodology. For instance, we've worked together with the universities of Delft and Groningen to create meaningful and statistically sound research methods. We're also constantly experimenting with new techniques to analyze and present data. We develop many of our visualizations and interactives ourselves. A big, upcoming innovation is a project to automatically generate news articles based on data sets, using text generation and artificial intelligence. For instance, we could create a unique article for every municipality in The Netherlands. For this project, we received a grant from the Google Digital News Initiative. RTL has hired a data scientist, who we are working with to develop this technology.

What was the impact of your project? How did you measure it?

Our data stories are some of the most popular articles on our websites, as well as being seen by more than a million viewers on television. But much more importantly: they have real societal impact. Many of our stories have led to parliamentary questions and debates, and incited policy changes. For instance: Our story on provincial roads has led to an increase in the budget to maintain these roads, our results on environmental noise levels were used by the Dutch public health service in their communication and the Dutch ministry of education showed an interest in our school quality methodology.
Our goal of creating useful news stories for our viewers and readers is demonstrably successful: our data tools generate a lot of interaction and shares on social media as well as positive feedback from users. It also has an impact on the rest of our newsroom: other journalists at RTL Nieuws see what's possible with data and are keen to learn how to use these techniques in their own reporting.

Source and methodology

We use a wide range of data sources. We like to use open data in new and innovative ways to find stories that no other media outlet has picked up. But we also work with businesses and academic institutions to find news using their data. If necessary, we collect data using scraping or Freedom of Information requests. Statistical reliability is very important to us. For this reason, we discuss our methods with experts, for instance at universities. We also aim to be as transparent as possible, providing readers with a detailed explanation of how we conducted a research, and sharing source data wherever possible. We also like to use the knowledge of insiders and our audience, for instance by using questionnaires. For example, this has helped us identify specific issues in our research on traffic safety near schools.

Technologies Used

Despite having only two data journalists, we use a broad range of techniques and technologies. This includes scraping and data analysis using Python, statistical analysis using SPSS, data cleaning and analysis using Excel, Tabula, Visidata and OpenRefine, developing interactive, reader-focused tools using JavaScript and jQuery, geographic analysis using QGIS and web mapping using Mapbox and Leaflet. We also use existing visualization tools such as Datawrapper and Flourish. Creating interactives for one of the largest news websites in the country means there are constraints on what's possible and usable across our sites and apps. We also deal with tight deadlines and quick turnarounds. That means we often have to be efficient and creative.

Project members

Wouter van Dijke, Jasper Bunskoek

Link

Additional links

Followers

Click Follow to keep up with the evolution of this project:
you will receive a notification anytime the project leader updates the project page.