Spatial Join (observations to line segment)

937
5
Jump to solution
04-18-2021 10:42 AM
AgnieszkaLisonek
New Contributor

Hi all, 

I am trying to join observations of the speed limit data captured by the car's GPS (>16k rows) to road segments (543 rows). 

I tried to use the spatial join tool and besides the "Closest" Match Option, which shows quite okay results, all the other options gave me <Null> data.

Is there any other function  I could use to join (map-matching) observations of the speed limit to the roads segment? 

Should I first set the projected coordinate system in order to perform this kind of features join? 

 

 

0 Kudos
1 Solution

Accepted Solutions
DanPatterson
MVP Esteemed Contributor

That is what the second link I posted covers.

If a location is equally close to two features, one is chosen at random.

Your option would be manually editing.

The Near tool also uses the same algorithm, so it is constrained by the data quality.

what percentage of your data produces what you call "errors"


... sort of retired...

View solution in original post

5 Replies
JayantaPoddar
MVP Esteemed Contributor

Since all the GPS points don't intersect with the Road segments, "CLOSEST" seems to be the best match. Had there been intersection between the road segments and GPS Points, "INTERSECT" would have been the better choice then.

Do you have any concern with the "CLOSEST" relationship?



Think Location
0 Kudos
AgnieszkaLisonek
New Contributor

I also thought that "CLOSEST" will be the best solution but sometimes it can generate error such as assign the speed limit sign to the wrong road segment, especially when close to the intersections.

I was wondering if there is any other method to get rid of such as mistakes. 

0 Kudos
DanPatterson
MVP Esteemed Contributor

From here

Spatial Join (Analysis)—ArcGIS Pro | Documentation

and here

How proximity tools calculate distance—ArcGIS Pro | Documentation

it is best to use a projected coordinate system (except Web Mercator, which is useless), to get the best results.

 


... sort of retired...
0 Kudos
DanPatterson
MVP Esteemed Contributor

That is what the second link I posted covers.

If a location is equally close to two features, one is chosen at random.

Your option would be manually editing.

The Near tool also uses the same algorithm, so it is constrained by the data quality.

what percentage of your data produces what you call "errors"


... sort of retired...
AgnieszkaLisonek
New Contributor

Thank you very much for your help! 

0 Kudos