Select to view content in your preferred language

Why over and undercounts spatial join points to line?

343
0
09-18-2020 01:01 PM
BarryFosberg
New Contributor III

I am seeking to join points to lines (Street segments). I have tried intersect, closest with and without search distances.

I have snapped the points to the lines and run without snapped points.

All joins are 1 to 1 Search distances vary when using Intercept. 

in the most recent case a join of 42,000 snapped points yielded a join count of over 66,000

Running the same points, pre-snapped means I must use a search distance. Results are  order of magnitude over counts or under counts and nothing in between.

Closest yielded join counts of 1 on every line or an 84,000 count total - against a 42,000 possible.

Using Pro 2.6

I made wild guesses at other possible join operation and get other strange results.

 

We think one cause of errors are points that lay at the juncture of two or more street segments. Why does the 1 to 1 rule not apply?

 

Suggestions? 

 

Update: using Arc desktop 10.5 and snapped points I get within 1% of the correct join count

 

Update to the update:

Using Arc Pro 2.8.

I am now using Summarize Incident Counts. That and a very large search radius has produced the right counts. Also I can count categories with a field. In my case this means I get a total count for each line segment and individual counts for each Shift and Crime Category based on an attribute field that has both.

A second solution is to join closest. 

Problem is solved.

0 Kudos
0 Replies