How to analyze signal intersection with crashes.

3917
13
Jump to solution
02-03-2015 07:15 AM
ElizabethPutnam
New Contributor II

I been given an assignment to group accidents that are near a intersection. I was given three shape files one for signal intersections, all intersections and accidents along the road for one county.

 

The problem that I am running into is overlapping buffers at the intersections. I am doing 250 feet buffers at the intersection. The reason for this is some intersections are too close to each other.

 

Buffer_overlap.JPG

 

The data that was given to me for intersections are broken into two or more dots one for traffic going left and the other for traffic going right at the same intersection. How can I get these two dots to join through out the whole county instead of doing it one by one.

 

Signals_Duplicate.JPG

If any one has done something similar to this and would be able to guide me it would be very helpful.

 

Thanks.

0 Kudos
1 Solution

Accepted Solutions
JoeBorgione
MVP Esteemed Contributor

I wonder if buffering your intersections isn't the way to go.  I just created two point feature classes: wrecks and intersections.  Then I performed a near between them.  Here is a snapshot of the augmented wrecks attribute table:

Look at wreck-points (OID) 1 & 2.  They are closest to the intersection with OID 2.  Clearly this is an extremely simplified analysis, but if I wanted to know which intersections have the most wrecks, I'd summarize this table on the NEAR_FID to get a count.  In this case intersection OID 10 only has one wreck associated with it and all the rest have two.  With a summary table you'll get a count of whatever you summarize on.  Then you could join/relate this information back to the intersections.  Make sense?

That should just about do it....

View solution in original post

13 Replies
williamcarr
Occasional Contributor II

Would it be possible to merge or append the buffers into a single polygon layer, then use that polygon layer to spatially join to the accident point layer?

ElizabethPutnam
New Contributor II

I did dissolve and then feature to point and did the buffer again and join my data. Thanks.

0 Kudos
AnthonyGiles
Frequent Contributor II

Elizabeth

if you turn the dissolve type to all in the buffer tool this will do the process in a single step.

regards

Anthony

ElizabethPutnam
New Contributor II

I looked into it and yes this saves a lot of steps. Thanks.

I'm having some data counts duplicated on my buffers that are overlapping. What could I do to avoid duplicate counts. I could make the buffer distance smaller but then the dots will not be included in either buffer.

Thanks.

0 Kudos
JoeBorgione
MVP Esteemed Contributor

Perhaps you can summarize your data and see if the id shows up more than once.

That should just about do it....
ElizabethPutnam
New Contributor II

This is what I am getting. The dots are being counted in each buffer and if they overlap they are counted in both buffers. Thanks for your help.

Double_Count.JPG

0 Kudos
JoeBorgione
MVP Esteemed Contributor

That's what I mean:  the buffers are goofing on you; if you do a straight up Near analysis, I'm thinking you should be good to go...

That should just about do it....
ElizabethPutnam
New Contributor II

Is there a tool where you can get the nearest/closest crashes that were not picked up by the buffer to be included in the nearest/closest buffer.

Thanks.

0 Kudos
JoeBorgione
MVP Esteemed Contributor

drum roll... Near Tool but it's only available that the highest license level.  I don't know what ESRI calls it but I call it ArcInfo. Also, the Point Distance tool may do it too...

That should just about do it....
0 Kudos