What struck me about what Steve Doig said was how powerful data journalism can be, but how it can also raise a whole new set of problems, and how even when it all works out, the story can have little impact.
He compared the situation in Florida before Hurricane Andrew with the situation in Louisiana before Hurricane Katrina. In Florida, the Miami Herald had been somewhat responsible for the shoddy construction of houses that fell apart in the storm. According to Doig, the paper had reported that some of the building standards were unnecessary, influencing the county commissioner to allow houses to be built that ultimately did not withstand the storm. Doig’s reporting after Andrew showed how lax home assessing had led to newer homes being destroyed at a much higher rate than older homes that had been built to the older, better standards.
But Hurricane Katrina brought up a different, yet still depressing, problem with journalism. The Times-Picayune had published a long series several years before Katrina about how sub-par buildings would be destroyed in a hurricane. They analyzed what strong winds would do to some of the buildings in the city, discussed how the city would flood because it is shaped like a bowl, and basically predicted the destruction that Katrina wrought. But according to Doig, no one listened, no one made improvements to the buildings or levees, and so the series ultimately had no effect and Katrina was an unprecedented disaster.
He also mentioned that even though he won a Pulitzer for his reporting about Andrew, and did many follow-ups to figure out how Florida should improve building standards, nothing much changed in Florida, either. He said that public officials pledged to change, but in the end, they didn’t do much.
I guess this isn’t a problem with data journalism per se, but it shows that even with mountains of data, fancy math, graphs, maps, and so on, making a difference even with something relatively simple like construction standards can be almost impossible. He said something like Florida has to have a hurricane every 25 years or else they forget that they need to take protective measures like building sturdy houses.
Another thing I noticed in his discussion was that the data itself can present problems — stuff like corrupt data gathered by volunteers who don’t care much about accuracy, or people’s names being misspelled and throwing off calculations. Also, problems such as campaigns not recording the gender of donors, creating a whole new problem of having to somehow analyze hundreds of thousands of names for gender. He described a complicated process he used to do that for an article. I can imagine a lot of other situations where you might have an amazing data set with tons of information, but it might be missing a key category, and it could be extremely hard to get the necessary information, especially if you have to add data to thousands of existing entries.
One thing that was encouraging about what he said was that despite the ubiquity of data in present-day journalism, most new reporters don’t have skills even in basic number-crunching like using Excel. That gave me a little hope that having these skills might help me get a job.