AP counted railroad ties, examined video to figure out Spanish train's speed
Five days before black box data from the deadly train crash in Santiago de Compostela, Spain, was released, the Associated Press reported the train was traveling "well above the speed limit." How did it report that with confidence?
AP Europe Editor Niko Price tells Poynter that Panagiotis Mouzakis, a producer in the news cooperative's London GraphicsBank, looked at CCTV video from the crash and figured that since each video frame represented 1/12th of a second, AP could figure out the train's speed if it could measure part of the distance the train travels in the video.
They picked two pylons in the video, Price said, and asked photographer Alfonso Bartolome Zafio, who was on the scene of the crash, to take a GPS reading in front of each. But police had cordoned them off, so he estimated the distance from a bridge above. Then, as a company memo relates, some newsroom magic occurred.
To back [Bartolome's estimate] up, Europe Desk editor Fisnik Abrashi used an AP photo picked up from a local newspaper to count how many cross ties there were on the tracks between the pylons, and editor Bob Barr, a railroad buff, researched the typical distance between ties on European railroads.
Each method gave AP a range of possible speeds: one was 89-119 mph; the other 96-112 mph. The speed limit was 50 mph, so AP was able to report far ahead of official announcements that the train was traveling about twice the speed limit.
"The preliminary investigation shows the driver braked from 119 mph to 95 mph in the moments before the crash," Mike Oreskes writes in the memo.
"We had a lot of people doing a little bit of work on it," Price said by phone. "It wasn't that everyone was running around trying to do this one thing for two hours." But, he notes, it couldn't have happened without people working across departments. Coming up with an estimate "ended up being a piece of our story," he said, "but it allowed us well ahead of anyone else to answer one of the key questions that we had at the time."
Also, he notes, the estimates "kind of disprove the old adage that a journalist can't do math."