0

I have a some point shapefiles that represent categories of folklore story, witch, mermaid etc. I also have some point shapefiles of landmarks like portal dolmens, ancient woods etc. Is there a tool in ArcToolbox to compare one group to another that would give me some result that says, 'The witch stories line up with the location of old churches more than ancient woods etc.

I've looked through the Spatial Statistics Tools and Spatial Analyst Tools but I can't find one that I think can do the job. Does anybody know if it is possible to do something like this in ArcMap?

nmtoken
  • 13,355
  • 5
  • 38
  • 87
Dakan_GIS
  • 51
  • 6
  • 1
    I think you're going to have to define what you mean by: "The witch stories line up with the location of old churches more than ancient woods etc. " to find the right tool and get a good answer. – GISHuman May 01 '18 at 19:56
  • I would start with a near analysis. Depending on how your data are structured, you may need to do it iteratively (or merge your shapefiles first). Then you can simply summarize by categories...again, it depends on how the data are structured. – Tom May 01 '18 at 20:47
  • 2
    Here is a possible related question, dealing with correlation between point datasets of various species of animals: https://gis.stackexchange.com/questions/83248/simple-way-to-compare-ranges-and-distributions-of-species-looking-for-correlati?noredirect=1&lq=1 – cmrRose May 01 '18 at 21:08

1 Answers1

1

You would have to probably do this in several steps to get statistics that would show correlations like you are describing. Depending on how many comparisons you want to make you may have to redo the analysis in several combinations.

I would start by using the Spatial Join tool under the Analysis Tools Toolbox, Overlay Toolset to join of the two feature classes together. The statistical results you can get will depend on which feature class is your Target FC and which is your Join FC and the Join Type you use. I would probably start by making the story category points the target and the landmarks the join and use the JOIN_ONE_TO_MANY option to assign every individual point with the landmark type it overlays.

Once you have the overlay, then you just have to keep doing different summaries using the data to see all of the relationships quantitatively. For example, you can use the Dissolve Tool or Summary Statistics Tool with the story category and landmark type as case fields. If you do Dissolve, be sure to get a count of one of the fields so that you can sort on the count to find the pair of types that get the highest number of hits. You could include the Join_FID in the case fields if you wanted to know the polygon with the highest hits for each story type.

To do a wide array of statistics without the need to create multiple outputs, you would need to learn Python. With Python it is possible to use cursors, dictionaries, lists, numpy, etc. to compile a wide array of statistics by just doing one or two passes on the overlay data and output them to several formats. But you would need to define the kinds of statistics you really are looking for and make an attempt at doing the coding yourself to get much help on a script designed for your needs.

Richard Fairhurst
  • 8,565
  • 1
  • 16
  • 19
  • Thank you Richard for taking the time to do such a detailed answer. I'll make a start on that and let you know how I get on. – Dakan_GIS May 04 '18 at 09:55