With Our Future Selves News21 is nominated for an Data Journalism Award. An interview with News21 journalist Michael Keller: ‘We developed a frame — Our Future Selves — and wrote a four part interactive that uses data to walk people through the next forty years of their lives — across population changes, health concerns and finances.’
What inspired you to make ‘Our Future Selves’?
We produced this story as part of a larger News21 project run out of the Columbia Journalism School called Brave Old World. News21 is a fellowship run with the Knight Foundation that aims to foster reporting on issues that have a bearing on the country as a whole. Our project was to tell stories about aging America and this interactive was one way we tried to bing people into that story. We partnered with the Washingon Post to publish this and other stories in our package.
Did you work by yourself or in a team?
I worked in a team with Emily Liedel and Jason Alcorn, two other News21 fellows at Columbia. We also had support from our two main editors Paula Span and Duy Linh Tu.
I think these projects are best when you have a team with whom you can throw ideas around and test concepts.
How did you get a hold on the data you needed?
We used five data sources with a wide range of difficulty in procurement. After identifying what data would drive the story and would give context to readers’ future lives, we started calling. The census provides the national demographic projections through 2050 and state level projections through 2030.
One interesting data set we wanted was mortality rates for men and women at every year of their life given any birth year. After spending a week trying to find a working number for the National Center for Health Statistics at the CDC I found the address, google mapped it to find the number of an adjacent building, got the right number and called. The main line had no receptionist however, only a dial-by-name directory. I blindly entered “AND” thinking someone by the name of Anderson would be in the directory. As it turned out, Robert Anderson is the chief of the Mortality Statistics Branch and he was very helpful throughout. He pointed us to the Social Security Administration who was able to send us a Cohort Mortality Table — used in part to measure Social Security Solvency — which was exactly what we needed.
The CDC only studies diseases but not the effect of those diseases across races, ages, or sex. Emily and I, especially Emily, computed these statistics ourselves with guidance from the Columbia stats department, using survey micro data compiled by the Minnesota Population Center.
Emily also spoke at-length with the Irban Institute to identify the areas that would be interesting to highlight for retirement finances. She was able to get data from them on retirement income and expenditures across race, sex, and marital status.
Which tools were used making this ‘Our Future Selves’?
We built this Flash and used Excel and TextMatefor data cleaning and construction of our XML data.
How did it take to make ‘Our Future Selves’?
News21 ran from June 1 to July 31 2011 and we finished it and the other stories we were working on by that deadline.
Were there any bumps in the road?
Our challenge was to make a story about aging interesting and relevant to people, especially younger people who don’t see themselves as a part of this aging trend. My generation knows that in 2050 we’re going to be senior citizens, and that personal choices and policy decisions made now will affect us then. But many of us still have a hard time thinking, I’m going to be older and I’m a part of this aging America trend even though I’m still young.
But the hardest part was deciding what data would best tell this story and then finding rich enough data that it would be applicable to a wide range of readers (I.e. we would have data across sexes, multiple races, ages etc). The health data was one example where we thought the CDC would have disease prevalence at least across age race and sex but they only compile these figures across one-dimension at a time. That left us looking for a new data source and collecting it ourselves from the http://ihis.us survey micro data. Also, when you’re dealing with data at this granular-level, you have to pay attention to sample size. We consulted with professors in the Columbia Department of Statistics to choose the right level of characteristics we could drill down to while making sure our margins of error were small. For instance, we initially wanted to include education level along with age race and sex but the samoe size in many combinations of these variables was too small.
Do you have a useful tip for starting data journalists?
Have a fun or challenging idea and try to build it. Code or knowledge from that project will most often be useful later. Keep tinkering because the technology is changing quickly. But also be weary of new tools that promise to replace your current workflow. They might not do everything you need them to do on deadline. Knowing the sure way you know how to get data cleaned or a vis programmed is often better than the fancy way.