Derailed Amtrak train sped into deadly crash curve

Derailed Amtrak train sped into deadly crash curve

Organisation: Al Jazeera America (United States)

Publication Date: 04/07/2016


Size of team/newsroom:large


This was a breaking news story a year in the making. In May of last year, an Amtrak train derailed on a curve outside of Philadelphia tragically killing five and injuring over 200. The crash was unusual for such a commonly trafficked corridor between Philadelphia and New York. The next morning, Al Jazeera America was able to publish the exact speed the train was traveling moments before it derailed: 106 miles per hour, over twice the speed limit for that stretch of track. We were able to do this so quickly because a year prior, I had started scraping the Amtrak "Track a train" map every five minutes and storing the data. The data provides real time location and speed of every train in the country. When we got to the newsroom, I located the records from before the crash and we were able to pinpoint its trajectory in an interactive and annotated map. As a followup the next day, we analyzed hundreds of train trips that went through the same curve to show that Amtrak 188 was an outlier: most trains were under or very near the 50mph speed limit. This analyze required analyzing gigabytes of historical data and using sophisticated geospatial analysis that had to be 100 percent bulletproof on a tight deadline.

What makes this project innovative? What was its impact?

The project was innovative for its use of proactively collecting newsworthy data and being able to analyze it on a tight deadline. We utilized the fairly new Turf.js Nodejs geospatial library, which helped us whittle down our national dataset of tens of thousands of files and filter them down to just those trains that passed through the same curve and that were, importantly, going the same direction. Personally what I like about the project is that one of the most compelling visuals is a histogram, which is a very simple but underused visualization type. In the spirit of openness, we published the Amtrak 188 data ( as well as the geo-bounding box we used in our analysis ( We also published a longer write-up sharing our process:

Technologies used for this project:

NodeJS, TurfJS, Underscore, Indian Ocean
Follow this project

Comments (0)

You have to be connected to contribute

You have to be connected to follow

Leave this project and no longer be informed about this project

By joining this project, you will be informed by email when an update or a new contribution is posted on the website.

Thank you for your active participation !

The GEN Community Team