Crash mapping

[Rauschpfeife]

Myles Smith (no doubt a cousin of mine) pointed out, in a sub-thread, that the Maine DOT offers a rather nifty crash mapping facility: 

https://www.maine.gov/mdot/mapviewer/

This version of the map shows pedestrian and cyclist crashes. The display is sub-optimal from our point of view, though, or so I think, and I did a bit of tinkering to improve on it, using DOT's exported data: 

https://www.smithbowen.net/portland_crashes/

The pop-up on each point tries to winnow through the considerable quantity of data in each DOT record and distill out the stuff that seems most of interest to us.  We probably aren't very interested in whether a moose was involved, for example, but this being Maine, there's a DOT Boolean field for it. 

-- 

Best, 

Michael Smith

anacharsis.substack.com

[Scsmedia]

Hey, it is a big deal when a Moose shows up in Portland.  It makes the local news.

Some of the state data is lacking and will never be updated, such as things like "no contributing action" and "Not stated"

I need to question this data point.  It is recent and says someone was seriously injured at the Casco Bay Ferry Terminal.  Was this the actual location or where the report was taken (would have to reach out to PPD to find out).  

Same happens with regular police reports taken at Police headquarters which make 55 Middle Street appear to a hive of criminal activity.  I notice there are not crash reports listing 55 Middle Street.

The Western Prom appears to be one of the safest areas of the city.  The four green number dots are not in the WP.  Not surprised to see Danforth and Vaughn to have three.  But only one crash on Danforth Street surprises me.

Steven Scharf

207-415-4662

SCSM...@aol.com

--
You received this message because you are subscribed to the Google Groups "Portland Bicycle and Pedestrian Advisory Committee" group.
To unsubscribe from this group and stop receiving emails from it, send an email to PBPAC+un...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/PBPAC/CAERgV99Uuf%2BM9rP11psHGj2aOD7eVFPv%3DiUYjHCqUNudFELaHg%40mail.gmail.com.

[Rauschpfeife]

Don't tell my landlord, but my roommate is a moose. Once I got here, I thought I should really get into the Maine thing. He doesn't get out much, though, and so far, he hasn't tangled with any drivers. 

I don't think we have the police report for the data point that Steven queried. The lat/long are what's in the DOT database. Those seem to come from the police report. 

The "no contributing action" and so on is also straight from the police report. There are four different fields, and it's difficult to display them in a way that's both perspicuous and compact. I'll give this some more thought. 

I don't know how the process of getting from police report to DOT database works. I hope it's not a matter of manual transcription. That's a recipe for trouble. The police report number and the DOT identifier are not the same, and there's no obvious way to derive the one from the other. 

One obvious statistical note: places where there are a lot of crashes are not necessarily more dangerous places, though they're undoubtedly places that need more attention. Ceteris paribus, you'd expect crashes to be proportional to traffic density, and crashes involving bikes and peds to be proportional to the *product* of car traffic and bike/ped traffic (interaction density). 

  

-- 

Best, 

Michael Smith

anacharsis.substack.com

To view this discussion visit https://groups.google.com/d/msgid/PBPAC/2003391587.3685704.1766467100074%40mail.yahoo.com.