GIS, mapping, crash reports vs. ASDM

Some notes on mapping using the latitude/longitude; and the ASDM (Adot Safety Data Mart) dataset.

Here is a detailed breakdown of a crash chosen more-or-less at random (I wanted to choose crashes that were *not* at intersections) that occurred 2012-10-12 at a driveway just east of 51st Ave on Indian School Rd. (if the link doesn’t work use 33.494971/-112.167771, the lat/long specified in ASDM). It is ADOT incident number 2672854, Phoenix file number 12001836231 (though it was listed as 201836231 in ASDM).

Notes on the ACR

The ACR (Arizona Crash Report) was written by hand.

The point of doing a nit-picky review of the ACR  is to see 1) how accurately it gets coded into ASDM, and 2) how it might be able to infer from only the ASDM dataset the type of crash that occurred.

A couple of besides-the-point issues: Calling this a hit and run is somewhat dubious; the bicyclist, thought to be a 10-13 year old boy,  “ran”, however it appears 1)no one was injured; though the bicyclist might have had a limp, and 2) there was “none” damage to the vehicle. The second aside is the driver apparently had no insurance; though the report doesn’t state anything more about that.

The geo-spacial entries are excellent: Onroad Indian School, Crossing feature 51st Ave, 338′ approx, east. Also the diagram was quite clear (they apparently denote feet with the “degree” symbol, a superscripted ‘O’?)

The mechanics of the collision are very clear; the bicyclist was EB on the N sidewalk (i.e. the “wrong” way) of Indian School Rd, when a driver leaving a private drive turning right collided with him.

The narrative doesn’t state that the bicyclist was on the sidewalk, only that the rider was E/B on the N/S of Indian School.  The NonMotoristLocation block does say SIDEWALK. And the diagram indicates the point of impact was 3′ N of the curbline. I’m guessing it was, indeed, on the sidewalk. The narrative should state that clearly.

In two places, the officer wrote up the actions of the bicyclist in a dubious manner; selecting “OTHER”. For Unit#1 (the bicyclist) violation/behavior block has written in “Rode on Wrong Side of RD” — riding on the sidewalk, in any direction, is *not* a violation. More confusingly for Unit#1 action wrote in “Riding Against TRF” — should have been GOING STRAIGHT AHEAD.

Block: Relation to Junction is marked as NOT JUNCTION; this is a major booboo in terms of when the data gets coded. The proper answer was DRIVEWAY. And then Type of intersection, and Trafficway Description were both slashed-through (as though they didn’t apply)

Block: direction of travel was listed as East/Southwest for bicyclist/motorist. I believe the proper answer (because it’s the direction “before 1st crash event”) is South. the SW apparently came from the fact that the motorist was egressing southward, and turning right, i.e. to the west.

Block: CollisionManner is ANGLE — that’s good and is the obvious choice, although for reasons noted here, cyclist collisions are sometimes arbitrarily listed as OTHER.

Other blocks of interest:

Lighting listed as DARK NOT LIGHTED; that area is almost certainly lighted.

Speaking of lighting; it’s probably actually relevant whether or not the bicyclist was lighted; however the report states nothing about that.

The ASDM data

You can view the entire data record for ADOT incident number 2672854,

The coding is, from what i can tell flawless; that is it reflects exactly what was written (by hand) in the report. Note that written in fields, e.g. the “other” fields, or the narrative, do NOT appear in ASDM.

The lat/long appears to be spot-on. The offset of 0.064 converts (in miles, apparently OffsetUnit=1 means miles) to the aforementioned 338 feet; though it is located with reference to the center of the nearest intersection (i.e. 51st and Indian School). It says the ValidLocationFlag is true, and this interesting-sounding piece of data that I don’t know what the heck it means: IntersectionRouteID = 11868

Given the directions, unit actions, and the NonMotoristLocation of sidewalk — someone would probably be able to correctly infer from the ASDM data alone (say, for example, if the ACR was unavailable) that this was a “counterflow” sidewalk cyclist colliding with a right-turning motorist at a driveway; though that would not necessarily be accurate due to things mentioned above such as and OTHER unit action for cyclist, and a SW (instead of S) for the motorist.

A sidenote: the individuals’ injury status was UNKNOWN, and NO INJURY — this translated into a “top level” (i.e. from the incident table) InjuryStatus of NO INJURY.

Here’s another one

There is an interesting inconsistency between the ACR and the ASDM. Adot incident #2693669, on 2012-12-18. This ACR appears to be some sort of computer-generated template. It is phoenix incident number 12002257862, which was FileNumber 122257862 in ASDM.

The NonNonMotoristLocation in asdm is NOT_REPORTED_255, however in the ACR it is IN ROADWAY(not in crosswalk or intersection).

The investigating officer apparently chose that location because he thought that was the most correct, there is a crosswalk there (map), but he felt the cyclist was riding parallel to, but outside of, it; which is depicted on the map in the report. In the narrative, it states: Pedalcyclist was S/Bound on Point Pkwy (Riding on the East side sidewalk), When she came off the sidewalk, entering the roadway as she was crossing Guadalupe Rd…”. Still the IN ROADWAY… designation doesn’t seem right, and probably ought to have been MARKED CROSSWALK AT INTERSECTION; but in any event it shows the troubles with that block’s (limited?) choices in general, and its use by law enforcement.phx122257862diagram

This odd region is an artifact of the way streets are designed around here with very large radii at intersections; the one line of the crosswalk that is closest to the intersection is actually several feet away from the intersection (the “prolongation” of the curb lines). The large radii also means that traffic will go around corners at higher speeds. When a crash occurs, it tends to be at higher speeds; leading to an increase in severity of injuries (see is-phoenix-safe for some more discussion). This region is a bit of a no-man’s-land, defying categorization, it’s not in the intersection, and it’s not within the two lines of a marked crosswalk.

In the ACR the JunctionRelation is INTERSECTION-RELATED, however in asdm it is INTERSECTION; what’s the “right” answer? It was clearly, technically, not in the intersection.

In any event; there are some inconsistencies between the ACR and ADSM as to location. In the ACR’s location block, the “from” box is checked (as opposed to the “at” box) on the crossing road; it probably should be “at”; it was, after all, an intersection collision. Furthermore, it says it’s 42 measured feet east. I don’t know if that’s supposed to mean 42 feet from the center of the intersection, or what — but in ASDM the offset is zero (zero, exactly). So, again, did Adot override this and discard the 42′ and simply put in zero? dunno.

Phoenix-area intersections tend to look not like this diagram, but rather more like the large-radius dashed line

So how, exactly does the asdm data get changed? At first I thought maybe there were some logical rules but that doesn’t seem like it could account for these changes; does somebody eyeball it, and override things that don’t seem to make sense?

Back to the Phoenix’s street design; to put it in some engineering-speak, from the ITE: “Larger radius corners increase the length of pedestrian crosswalks, and increase vehicular turning speeds. In designing walkable urban thoroughfares, the smallest practical curb-return radii are used to shorten the length of the pedestrian crosswalks”