By the numbers: Wrangling data at the 2015 Journalism Interactive conference
Katie Park, graphics editor at The Washington Post, and Ryan Murphy, lead developer for news apps at The Texas Tribune shared their insights on data-driven storytelling and data-driven journalism education at the 2015 Journalism Interactive Conference.
- Inside the Swag Bag of the Journalism Interactive conference that I helped organizing. #jiconf https://instagram.com/p/13z8RRlieS/
- Norman Lewis, associate professor at the University of Florida led the panel.
“Understanding [graphics] is just another way to tell story”, said Murphy. “The goal is always the same as to deliver the information in the most effective way.” Here are five ideas and suggestions Park and Murphy shared about data-driven journalism:
- Park and Murphy suggested four categories of data-driven storytelling projects.
· Narrative/analysis projects that provide context or a narrative within large datasets. Park said the reporters and editors prefer to find anecdotes to personalize the stories, yet it is also important to show the audiences a broader picture, and this is where the data comes in. NPR’s Walmart’s urban spread story and theGuardian’s measles story are both good examples.
· Exploratory projects in which users may find patterns that interest them. Examples include The Texas Tribune’s salaries explorer story and the New York Times’s 2,000 calories story. 
Eating the drought: How much water goes into your meal?California's crippling drought has prompted conservation efforts, such as replacing grass lawns and minding how long you leave the tap water running. But what about the food on your plate? Agriculture uses 80% of California's water supply, and producing what you eat can require a surprising amount of water.
· Utility projects in which users can find out data that applies to themselves specifically. Examples include The Washington Post’s DC school boundaries story and the Texas Tribune’s vaccine exemptions story.
Government Salaries Explorer | The Texas TribuneUse The Texas Tribune's Government Salaries Explorer to review the compensation of hundreds of thousands of state and municipal employees in Texas.
Watch how the measles outbreak spreads when kids get vaccinated - and when they don'tIf you take 10 communities and run a simulation, it's easy to see why we need as many members of the 'herd' as possible to get vaccines - before it's too late
· Participatory projects that allow users to provide their own data. This kind of project makes the newsrooms go beyond the information they have, and invite the audiences to contribute back to their projects. Both The New York Times’s story on reaction to Bin Laden’s death and NPR’s playgrounds story give the users a chance to update the information themselves.- They provided a shared Google Doc with lots of resources:
- Being part of the conversation:
Park and Murphy said the graphic editors have to know what the story is about before they make it visual to better deliver the information.
The traditional way of getting a story idea, writing the story and then adding visual elements at the end may limit the scope of additiona lcontent and prevent reporters from thinking about the articles in different ways, Park said. - Being mindful of mobile platforms:
As the number of mobile users increases significantly, publishing on mobile platforms becomes a concern. Almost all the Washington Post’s projects are responsive, Park said.
Murphy also suggested that running tests on different platforms and browsers to see whether they work well is necessary.
- "Interviewing your data" -- what would you ask your data if you treated it like an interview subject? - @katiepark #jiconf
- Viewing data as sources:
To set up “stepping stones,” Park prefers to keep a basic list of statistics that generated from her datasets, such as the low and high values, the average and the median values.
“This is the concept in data journalism called interviewing your data,” said Park. She suggested reporters approach the data like interviewing their sources and think of what questions can be asked.
Murphy said bias may also exist in the dataset, so the reporters need to examine the dataset carefully and, if necessary, request help from experts to verify the database.
He said it is also useful to build a source list and constantly request databases from different agencies. - Every data project needs a narrative. Don't just put a chart on the side of a story! - @MedillSchool alumna @katiepark #jiconf
.@katiepark talks up web literacy as key skill, says not every journo needs to know how to code, but should know how the web WORKS #jiconf
@katiepark says you don't need to be a stats expert--but how does your data sort? How calculate basic avg? Range? Standard dev? #jiconf- Importance of spreadsheet literacy:
Park and Murphy both considered spreadsheet literacy as a base skill to reporters.
If reporters are able to look through and understand the dataset, and do some basic analysis by computing the numbers, it saves them alot of time, Park said.
She also suggested reporters keep a data diary to help them keep tracking on their steps and recall wrong instructions.
“The communication [between graphic editors and reporters who can work with spreadsheet] is so much easier,” Murphy said, as this skill helps reporters build connections with them.
Below are some resources that Park and Murphy introduced:
For further study:
· Robert Kosara’s data viz basics
· Flowing Data Learning
· Matt Ericson’s “When Maps Shouldn’t Be Maps”
· Journalist’s guide to datasets
· Courseoutline/projects from Derek Willis’s 2012 Digital Frameworks class
For inspiration:
· thedailyviz.com/
· Graphics, data and news apps on Twitter
