Over the last year the Data Journalism team has taken a key role in visual storytelling at the Telegraph, collaborating with New Formats, Graphics and Editorial Development teams to improve and expand our storytelling output. Our work includes innovative special projects, exclusive news reporting and personalised or interactive content to engage our subscribed users. This year, our team has been particularly focused on leading the way with The Telegraph’s new subscription-first model, and we have ensured that data journalism is at the centre of a sustainable business model for the publication.
Our bespoke analyses and visualisations provide our community of subscribers with a deeper and richer digital experience – one worth signing up for to get data-led content that can be interactive, personalised or original. Such an audience – often a middle-aged UK audience who are particularly interested in politics, personal finance and sport – benefit from data journalism and visuals that are insightful but succinct. This year has also seen us deliver more exclusive news stories than ever before, producing splashes and page leads with data revealed through investigative processes like freedom of information and targeted scraping.
The team consists Ashley Kirk (senior data journalist), Patrick Scott (senior data journalist), Sarah Newey (data journalist, with a global health security patch) and Joshua Wilston (data journalism apprentice). We work in conjunction with a variety of other teams to come up with new and effective ways of communicating information to our readers. This could include producing interactives where a story warrants it, personalising stories to tell readers how, for example, a story on unsolved crimes impacts their local community specifically.
What makes this project innovative?
A true Telegraph innovation lies in the way we have adapted our data journalism approach to support the publication’s business model. Data-led tools and analysis now sit behind paywalls in open articles, allowing a two-tier approach to reader engagement. Nobody else is doing this, where we are introducing paywalls into non-traditional storytelling formats. We have enriched stories through personalised premium (i.e. paywalled) embeds, that have added a layer of detail only available to those who are logged-in. We have experimented with different paywall strategies to good effect, attracting praise from the wider industry. For example, others in the data journalism community commented on how our Manchester City piece was the first project they’d seen to experiment with combining registration walls with the scrollytelling format. As well as the interactive, visual storytelling content, we have also delivered dozens of exclusive scoops through our reporting. Our team has used hundreds of targeted freedom of information requests to deliver several stories such as one that revealed police forces' average emergency response times and found that every force had seen response times increase. This produced a page lead for the newspaper as well as a visually-led article for the website. The visualisation, which took a long time to craft due to the colour scheme reflecting gradually increasing target times, was commended by visualisation expert Andy Kirk. Further innovations have taken place in special projects. To set ourselves apart in our coverage of the 2018 World Cup, we produced an interactive game which allowed readers to pick what they believed were the most important factors in deciding a football match, and then played out every game based on these weightings to decide who had the best chance of winning. This was innovative as we put the controls firmly in the hands of the user. Rather than force our perspective on to our community, we opened up the data and allowed them to decide what was the most important factor in football.
What was the impact of your project? How did you measure it?
Our data journalism work has contributed to the success of the organisation’s wider subscriptions strategy. While others have struggled to make paywalls work, we have grown the number of registered users - repeat, signed-in readers - by more than 350% YoY. With our Premium subscriptions enjoying YoY growth of 55%, this model supports a commercially viable future. Over the past year, The Telegraph’s business model has evolved from a registration- to subscription-first approach becoming our key business priority. This is something we have paid a lot of attention to and have made sure that data-driven stories perform well in this regard by adding value. Indeed, some of our pieces in the past year have been among the best performing on the website. Our Buy-to-Let tracker has, for example, been very successful in this regard, showcasing exclusive data from Hometrack in interactive visualisations alongside expert advice from journalists from our Business desk. The article is one of the top five editorial pieces from the past year in terms of attracting subscriptions. As well as this, our World Cup winners simulation game was among The Telegraph’s top five stories for registrations for the World Cup, a significant achievement for a new form of experimentation for us. As well as delivering solid innovative content that sustains our subscriber community, we have also continued to break exclusive news stories. Our team’s stories often lead both The Telegraph’s website and newspaper. In print, freedom of information-led stories such as “Soaring numbers of patients sent home from hospital at night" have made significant page leads in The Telegraph newspaper, demonstrating how we have helped make data journalism a driving force for news reporting. Such print pieces inevitably make solid, graphics-led news reports online as well, usually gaining thousands of subscriber views and high dwell times from our engaged community.
Source and methodology
We use data from a range of different sources. These can be official government publications, data releases from other organisations (e.g. polling companies), but can also be self-created through investigative methods such as scraping or freedom of information requests. For example, using Freedom of Information requests at a national level alongside interviews, we were able to build a dataset which revealed that over-stretched UK hospitals were being forced to send home increasing numbers in the middle of the night. This was a challenging piece as we had to wrangle together 150 inconsistent freedom of information responses, but data cleaning processes help this and the result was rewarding. Our analyses are often simple statistical exercises (usually in R - but Excel and Google Sheets are great too!) but can become more complex. In the past year we have used machine learning, natural language processing and linear modelling to come up with stories. For example, our World Cup forecasting game used team and player performance data to create a linear model for predicting the tournament.
Ashley Kirk, Patrick Scott, Sarah Newey, Josh Wilson