This project seeks to exploit a kind of natural experiment in the District of Columbia by examining crime in close proximity to street lights before and after they have been repaired.
The simplest analysis we can conceive is simply examining crime the week before and the week after a repair. This will require (1) goecoding the crime data; (2) geocoding lighting; (3) drawing some reasonable polygon -- perhaps a circle or perhaps a city block) around each light; (4) joining the lighting polygons with the crime data surrounding each light for the week prior to and following the repair. More sophisticated analyses will include standard statististical analyses and, possibly, machine learning analyses to predict where improved or upgraded lighting could do the most good.
Happily, DC maintains a great deal of open data about streetlight locations and crimes known to MPD.
DDOT maintains both a detailed inventory of steetlights in the District and records of repairs. Repair data spans two data sources: (1) the current CityWorks workorder management data system that DC adopted in 2016, and (2) the now-retired iSlims work order management system that managed work orders prior to Cityworks. CityWorks data is available via a public API, and DDOT has generously provided us with access to data from the retired iSlims system.
-
Location and feature information & geojson
-
CityWorks service request information & geojson
-
CityWorks work order information & geojson
-
iSlims Work Order Data: available in the data subfolder. (We know, we know, but it's not big and until we get it up on the open data portal, this is just the easiest thing to do.) Please use the pickled data files. They are the most complete and up-to-date.
Public crime data, although exluding some offenses such as sexual assaults, is fairly extensive and includes many common offenses that one would expect to be influenced by lighting if lighting were to have an effect (theft from auto, robbery, assault, etc.).
-
2017 information & geojson
-
2016 information & geojson
-
2015 information & geojson
-
2014 information & geojson
-
2013 information & geojson
-
2012 information & geojson
-
2011 information & geojson
-
2010 information & geojson