I did a project on this some time ago. Columbia gas was one of the worst offenders for sheer volume of leaks. Not only is it normal to leave leaks unfixed, freq

I did a project on this some time ago. Columbia gas was one of the worst offenders for sheer volume of leaks. Not only is it normal to leave leaks unfixed, frequently they disappear from the public record without any recorded nearby fixes. It'd be interesting to see whether there's a correlation between explosion and leak location. Github: https://github.com/RogerTangos/LostLeaks/tree/master/data https://github.com/RogerTangos/LostLeaks/tree/master/data Overview (just for the Boston Area. I didn't do the analysis for Columbia gas.) http://lostleaks.csail.mit.edu http://lostleaks.csail.mit.edu I'm happy to lend a hand for the Columbia 2016-17, and 2017-18 data if others want to join in.