topic Re: Calculating averages using point data: Desperate for assistance!!!! in ArcGIS Spatial Analyst Questions

Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Wed, 21 Nov 2012 00:30:54 GMT

Hi,

I currently have a global map shapefile, superimposed by 8192 data points that show temperature changes (per degree Celsius of global warming) and degrees of latitude/longitude at those locations on the globe. So, for instance, an individual data point would say (when using the "Identify" function) at 180 West, 45 degrees North, the temperature change per degree Celsius global warming is 0.8 Celsius. Each of the 8192 points contain this sort of data.

What I am attempting to do is simply calculate various averages for different regions by somehow combining the necessary points over those regions. For example, if I would like to calculate the average temperature change in the United States, how could I do this using the point data that covers the United States?

Any assistance would be greatly appreciated!

Patiently awaiting your response(s),

~Trav.~

Re: Calculating averages using point data: Desperate for assistance!!!!

Anonymous User — Wed, 21 Nov 2012 01:06:54 GMT

If you simply want to find the averages in a particular region, you can do a spatial join of your Boundaries file to the point file. Once you have done this you will have the region name for all points and you can then get the average by region many ways:

1. right click the region name field and choose "Summarize" and select "MEAN" for the temperature field
2. use the summary statistics tool to do the same. Use the regions field as the case field and the temp with "MEAN" as the Statistics field.

http://resources.arcgis.com/en/help/main/10.1/index.html#//00080000001z000000

Re: Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Wed, 21 Nov 2012 03:06:38 GMT

Hi Caleb,

Thank you so much for your good assistance on this! I just wanted to know, however, if there are simple procedures to follow in conducting a spatial join?

Thanks again,

~Trav.~

Re: Calculating averages using point data: Desperate for assistance!!!!

RamB — Wed, 21 Nov 2012 06:24:20 GMT

Hi,

first make a polygon layer that has a polygon for each region you want a summary for.
second, do a spatial join to the polygon layer using the 'one to many' option (polygon to many points) with the point file and do not forget to check the option to calculate average value for a user defined field (temperature change).

after you do that, open your polygon attribute table and you have average temperature change for each polygon(region)

regards,

Re: Calculating averages using point data: Desperate for assistance!!!!

Anonymous User — Wed, 21 Nov 2012 12:19:24 GMT

Here is the link showing how to use the spatial join tool. It is located in the Analysis Tools toolbox under "Overlay". I believe the summary statistics is also in the Analysis tools.

http://resources.arcgis.com/en/help/main/10.1/index.html#//00080000000q000000

As Srirama suggested, use the "One to Many" join operation and make the temperature points the "target features" and your regions as the "join features". Once you have done this you can use one of the two methods I showed to get the average temp change.

Re: Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Wed, 21 Nov 2012 19:20:09 GMT

Thank you, Caleb and Srirama, for your good help on this!

I will attempt what you both had suggested, but I guess that my only question is would the "Spatial Join" function work for the desired regions on a global map layer, or do I need to separately create other layers with the regions of interest? For example, if I were interested in the U.S., could I use the "Spatial Join" function directly for the global/world map layer that I already have, or do I need to create a U.S. layer separately, and then use "Spatial Join"?

Also, how may I create separate "polygons" for specific regions?

Many thanks, again,

~Trav.~

EDIT: I have attached a map to facilitate my question.[ATTACH=CONFIG]19463[/ATTACH]

Re: Calculating averages using point data: Desperate for assistance!!!!

Anonymous User — Wed, 21 Nov 2012 19:26:38 GMT

Thank you, Caleb and Srirama, for your good help on this!

I will attempt what you both had suggested, but I guess that my only question is would the "Spatial Join" function work for the desired regions on a global map layer, or do I need to separately create other layers with the regions of interest? For example, if I were interested in the U.S. could I use the "Spatial Join" function directly for the global/world map layer that I already have, or do I need to create a U.S. layer separately, and then use "Spatial Join"?

Many thanks, again,

~Trav.~

The spatial join will work with all regions in whatever layer the boundaries are coming from. For example, if you have a feature class that has ALL the countries and/or continents and you spatially join those polygon features to the temperature point features will give you a result of all the point features with a field in the table that tells you which region each point is located in (plus the other fields in the boundaries table). So you can use the "Region Name" field as the case field for the summary statistics tool and select the "MEAN" statistic for the temperature field. This will give you the mean temperature between all the points within each region. If you just want to do individual regions, you can just select a region by its name and get the Mean for just those selected points.

Re: Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Wed, 21 Nov 2012 20:11:25 GMT

Hi Caleb,

Okay, that sounds fine, thank you!

Just to be safe, though, I have included an image attachment in my previous message showing the map (and superimposed data points) that I am using. Did you happen to see it above? If so, does this map fit your explanation? You can enlarge the image to see the details for the world map and what is included. 🙂

Thanks again,

~Trav.~

Re: Calculating averages using point data: Desperate for assistance!!!!

RamB — Thu, 22 Nov 2012 05:10:19 GMT

Hi,

you making it too complex while it is not. 🙂 It is a simple right click operation.

just use the following option on spatial join on polygon (join feature) with the point layer (target feature)

1. join operation = one to many
2. use the merge rule mean on your temperature field
3. match option = completely within.

It is solved. Open your polygon layer attribute table and you have mean temperatures for each polygon (each row).

🙂

Re: Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Thu, 22 Nov 2012 20:38:02 GMT

Hi Srirama,

Thank you for your assistance on this, and I apologize if I am a little too meticulous! I will duly proceed with your (and Caleb's) instructions.

Just to be safe, though, these procedures would apply to the world map that I have shown above (all of the layers that I am using are found in that image at the end of post #6, if it helps), correct? Sorry about the ongoing questions, but I would just like to make sure that I am on the right track. 🙂

Thank you for all of your help and patience.

~Trav.~

Re: Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Tue, 27 Nov 2012 19:47:25 GMT

Hi Caleb/Srirama,

Okay, I've done as was instructed with respect to "Spatial Join" and successfully acquired a new attribute table. I've tried joining the temperature point data (Target Feature) and "country name" or "Continent" (so, two attempts), based on the world map that I have (see attached file above) - however, under "Continent", for example, I noticed that for many rows, they say "<Null>". Some rows actually show the actual continents, but most are nulls. Is there a reason for that?

Now, if I want to simply acquire the annual temperature means for, say, Antarctica, how may I do this using the attribute table? I understand that you had mentioned right-clicking on the "Continent" column, for example, and then select "Summarize" in the drop down menu. However, how may I only select the Antarctica rows, and then acquire an average based on those selected rows?

Thanks again, and I apologize for the questions!

~Trav.~

Edit: Bill, thank you for your input. Would it be possible to show me a detailed procedure in doing this in ArcGIS? If so, I would greatly appreciate it!

Re: Calculating averages using point data: Desperate for assistance!!!!

Anonymous User — Wed, 28 Nov 2012 01:05:24 GMT

Those null values are most likely coming from the points that are not overlapping with the continents (aka the points over the oceans). Since the continents do not intersect the points over the oceans, there is no continent value for those rows (thus giving you null values). You can still summarize the table by the continents field though to find your averages.

Also, Bill made a great point and weighted averages would probably be a good idea.

Re: Calculating averages using point data: Desperate for assistance!!!!

TravMoore — Wed, 28 Nov 2012 01:21:03 GMT

Hi Caleb/Bill,

I believe that you may be correct about the nulls representing mostly oceans, or bodies of water. However, I noticed that there aren't a great deal of rows covering the continents (for example, the number of rows covering South America is less than the actual number of points covering that continent). Antarctica, though, has a comparatively large amount of rows. Could there be a reason as to why?

Now, if I wanted to select only the Antarctica rows in the resultant table (from the spatial join), for instance, how may I do this to calculate the annual average for Antarctica?

As for Bill's suggestion, this had also been stated to me previously, as well. However, I have absolutely no idea where to begin in calculating weighted averages. I took a look at the link that Bill provided, but I quickly became lost, unfortunately. Is there a way to obtain detailed step-by-step in order to compute weighted averages? I am really interested in going through with this!

Many thanks,

~Trav.~