Jump to content

 
Photo

overlapping point data

- - - - -

  • Please log in to reply
6 replies to this topic

#1
kfox

kfox

    Newbie

  • New Member
  • Pip
  • 2 posts
  • United States

Hi CartoTalk,

I have a question about how to deal with overlapping point data in ArcMAP 10. I have a large data set (over 50, 000 points) for doctors in the 50 US states. As you might imagine, these points (represented by dots) overlap in many places--sometimes directly on top of each other, making it impossible to see how many there are AND sometimes because they are just so many in, say, a 5-mile area that they just stack on top of each other. We want to be able to balance seeing all dots while maintaining (visually) their spatial reference. For now, we are trying to do this ONE-to-ONE, or one doctor=one dot. We may have to try working with "Collected Events" but that is not necessarily an option right now.

I have tried to use the "Disperse Markers" function by setting this to EXPAND (0 miles) but it spreads the dots far too much and we lose seeing where that point really is on the map. Some have recommended using dots as labels and setting things up there but I can't seem to finesse this to keep labels where I want them.

I realize that with the clusters of 500+ we will not be able to do much. But the cluster/overlap of 15 that looks like 3 is a real problem.

Any ideas?

Best,
Kevin

#2
David Medeiros

David Medeiros

    Hall of Fame

  • Validated Member
  • PipPipPipPipPipPipPip
  • 1,082 posts
  • Gender:Male
  • Location:Redwood City CA
  • Interests:Cartography, wood working, wooden boats, fishing, camping, overland travel, exploring.
  • United States

This is really just a map scale and extent issue. At a national scale your data are just going to start running on top of each other where they are close together. Rescaling the view will eventually separate the dots but that won't satisfy an overall shot of the distribution or territory. I'd look at using an Integrate and Collect Events process to consolidate points and classify by counts.

GIS Reference and Instruction Specialist, Stanford Geospatial Center.

 

www.mapbliss.com

 


#3
kfox

kfox

    Newbie

  • New Member
  • Pip
  • 2 posts
  • United States

This is really just a map scale and extent issue. At a national scale your data are just going to start running on top of each other where they are close together. Rescaling the view will eventually separate the dots but that won't satisfy an overall shot of the distribution or territory. I'd look at using an Integrate and Collect Events process to consolidate points and classify by counts.



Thanks David. I hear what you are saying for the national scale. We just have to accept that. When we want to map at the state level we can get 400-500 points lumped in one county. This, too, is acceptable. But it is in the county with say 15 overlapping points where I just want to be able to spread those points out just enough to see them all. And do this without messing with the larger metropolitan areas.

I agree about the Collect Events but everyone here still wants to try ONE=ONE.

Best,
Kevin

#4
frax

frax

    Hall of Fame

  • Associate Admin
  • PipPipPipPipPipPipPip
  • 2,304 posts
  • Gender:Male
  • Location:Stockholm, Sweden
  • Interests:music, hiking, friends, nature, photography, traveling. and maps!
  • Sweden

I think you should consider using some kind of heat map or coropleth map for the small-scale views.

Here is an interesting article about using hex-polygons for a heatmap (the latest gimmick...)
http://blogs.esri.co...scale-web-maps/
Hugo Ahlenius
Nordpil - custom maps and GIS
http://nordpil.com/
Twitter

#5
jrat

jrat

    Master Contributor

  • Validated Member
  • PipPipPipPip
  • 143 posts
  • Gender:Male
  • Location:Hagerstown, MD
  • United States

What about using nothing as your symbol but select a graphic as a label and then use the tools in the maplex placement to make them not overlap. Just make sure that the label engine doesn't drop any off.

#6
Craig Jr

Craig Jr

    Contributor

  • Validated Member
  • PipPip
  • 15 posts
  • Gender:Male
  • Location:Cold Spring, NY
  • United States

A suggestion that I have at the National Scale would be to summarize the points to the Counties to get the number value of doctors per county. Convert the County polygon to a center point and display the number of doctors per County at the National scale as a graduated symbol and label areas that have greater counts using a sql query in label expressions.
Or show a thematic map using the County polygons after you aggregate the point data.

You could also leave the points in their proper location and label the total count of doctors per County on the map.

Inserts showing areas with high clustering of doctors?

Good Luck!

#7
mfarmer

mfarmer

    Contributor

  • Validated Member
  • PipPip
  • 38 posts
  • United States

The heat map is a good idea. Another suggestion is to determine a geographic level (Block Group, Census Tract, County, etc.), to which you can spatially join your points. Then you could make a dot density map with the resulting data set.




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users

-->