Data Therapy for Journalists
How do we best tell a data-driven story? What techniques should we use for presenting our data? As part of my Data Therapy project I've been helping community groups tell their data stories creatively for 5 years. Now journalists are asking me the same questions. This post explores some of the techniques I see emerging for journalists to tell their data-driven stories. I've bounced these ideas off the Boston Hacks & Hackers and the PenPlusBytes Bootcamp for student journalists, and I welcome your thoughts, suggestions and feedback.
Data-driven journalism is not a new thing. Journalism has a rich history and tradition of data collection, aggregation, validation and filtering. That said, I see three things changing the landscape right now:
- The data has shifted from mostly qualitative to more quantitative;
- The sheer volume of data we have access to has increased significantly;
- There are more and more tools available to tell stories, leading to experimentation with new techniques.
These three trends have forced journalists, editors and the public to rethink the skills required to effectively tell stories. There are now people helping journalists get better at statistics to find stories in data. There are now people working with journalists to write code and build news apps. There are now people partnering with journalists to mine data to find stories. However, there aren't many people helping journalists figure out how to best to tell their data-driven stories. I'm trying to address that by explicitly listing specific techniques and their strengths.
Here are a handful of techniques I see emerging, each with an example. I hope these contribute to a larger conversation about patterns available to journalists for telling data-driven stories.
Let your Audience Explore
Present data to the audience in a way that lets them interactively explore to find stories they relate to.
The New York Times created an interactive graph to accompany their story about how people spend their day (based on the American Time Use Survey). They include filters (options you can click on to view subsets of the data) as an invitation to the reader to explore the dataset as it might relate to their own daily life. Creating exploratory interactive pieces can engage curious audiences. The field of data visualization is young, but mature enough to have a few standard tricks. One of these is to ensure that people looking at a large dataset have a way to project themselves into it, or compare themselves to it.
Explain with Pictures
Create an explanatory graphic to accompany your story.
Nigel Holmes has been creating "explanation graphics" for decades (we're calling these infographics now). Many of his graphics use cartoon techniques to explain data-driven stories. Graphic depictions of data stories are sometimes lampooned by the likes of Edward Tufte and data-density die-hards, but they are engaging and can be playful. Scott McCloud addresses many of the issues around these types of explanations wonderfully in his book Understanding Comics. This gives you an opportunity to walk readers visually through your story.
Create Striking images
Make a creative representation of your data to introduce your story.
Peter Orntoft took Danish survey data and created scenes to photograph that represented the survey data. For survey data about the wearing of religious symbols in public, he photographed religious symbols he had modified into physical representations of the survey results. Letting yourself have a little fun with representing your data can be an effective way to disarm your readers. It can break down assumptions they might have about the data or topic at hand. Plus, it's fun!
Share Your Investigation
Dig through large data sets to discover stories that you can tell in traditional "newspaper language".
ProPublica won a Pulitzer prize for their Magnetar Trade story. They sifted through massive datasets of financial records and found a compelling narrative about how this one company plotted to keep the housing bubble going, contributing to the financial crisis. This is one of the more traditional techniques on my list, building on the long tradition of investigative journalism. The key development I see here is sharing the journey through the data that led to the story. The idea that you can be more open about your process can lead to new relationships with your audience, engaging them as co-tellers of the data-driven story.
Advice on Picking a Technique
Should you build an exploratory news app? Should your story have an explanatory visualization? These are the hard questions. There are no firm rules about when to use each technique. I always suggest basing your choice on your audience and your goals. If you know your readers don't have lots of time, an exploratory web app isn't going to work for them. If you want to grab people's attention, an exciting image can quickly grab your reader's interest and give you time to hook them in. If your data tells a complex story about an issue unfamiliar to your readers, an explanatory graphic might help. These are just a few ideas about how to choose between techniques. Often you want to use more than one. What do you think? Is this type of list handy when you sit down to do your data-driven storytelling?