# Spatial analysis - globe and distance

arcgis raster spatial analysis sphere gis population demography

6 replies to this topic

### #1 frax Posted 27 November 2013 - 02:59 AM

frax

Hall of Fame

• 2,336 posts
• Gender:Male
• Location:Stockholm, Sweden
• Interests:music, hiking, friends, nature, photography, traveling. and maps!
• Sweden

I was just thinking about an interesting thing... Have you seen this map that made the rounds in social media a year ago or so: http://io9.com/more-...ide-t-493103044

I was thinking about how to repeat that analysis, but for an arbitrary point, especially close to the -180/180 meridian, or closer to the poles. Let's say one picks a point in Chukotka (far Siberia), Hawaii or Alaska, how could one create this circle (easily) in a software like ArcGIS.

Let's say I have a population dataset with absolute population in each grid cell for a 1 deg raster. Doing the actual circle analysis is not so hard, those tools are built into ArcGIS (spatial analysis), but handling the edges of the raster is different.

This might be one case where it would be easier to maybe calculate this using spherical coordinates.

For ArcGIS, I am thinking a bit about shifting the dataset so that the arbitrary point is in the middle (0,0), do the analysis, then shift back... Not sure how to do the shift though!

Hugo Ahlenius
Nordpil - custom maps and GIS
http://nordpil.com/

### #2 Hans van der Maarel Posted 27 November 2013 - 03:42 AM

Hans van der Maarel

CartoTalk Editor-in-Chief

• 3,944 posts
• Gender:Male
• Location:The Netherlands
• Interests:Cartography, GIS, history, popular science, music.
• Netherlands

Shift to a different central meridian is not that difficult. Say you want to center on +150 degrees

1. Cut the data at -30 degrees (150 - 180).
2. Shift the left portion of the data 360 degrees to the right so the map runs from -30 to +330 degrees.
3. Shift the entire dataset 150 degrees to the left.

Just tried it out with FME and it works. Can supply the workbench if anybody is interested.

The other part, shifting to a different central parallel might be more tricky. I've looked into that a long time ago for an aborted globe project. I'll see if I can dig that up.

Hans van der Maarel - Cartotalk Editor
Red Geographics

### #3 frax Posted 27 November 2013 - 08:53 AM

frax

Hall of Fame

• 2,336 posts
• Gender:Male
• Location:Stockholm, Sweden
• Interests:music, hiking, friends, nature, photography, traveling. and maps!
• Sweden

Yup - the x-shift is easy in ArcGIS as well, one can just switch to a projection with that meridian defined as center.

Hugo Ahlenius
Nordpil - custom maps and GIS
http://nordpil.com/

### #4 Strebe Posted 27 November 2013 - 11:30 AM

Strebe

Key Contributor

• Validated Member
• 92 posts
• Gender:Male
• Interests:Map projections. Snobby chocolate. Science in general.
• United States

I'm confused about the question here. The goal is to put a circle on the map, not on the globe, is it not? If so, why would you shift the central meridian or parallel to do your calculations? If you do that, then you'll distort your circle when you shift back

--daan Strebe

### #5 DaveB Posted 27 November 2013 - 12:14 PM

DaveB

Hall of Fame

• Validated Member
• 1,074 posts
• Gender:Not Telling
• Location:Redlands, CA
• United States

never mind

Dave Barnes
Esri
Product Engineer
Map Geek

### #6 David Medeiros Posted 27 November 2013 - 12:57 PM

David Medeiros

Hall of Fame

• Validated Member
• 1,131 posts
• Gender:Male
• Location:Redwood City CA
• Interests:Cartography, wood working, wooden boats, fishing, camping, overland travel, exploring.
• United States

Yup - the x-shift is easy in ArcGIS as well, one can just switch to a projection with that meridian defined as center.

You can re define the central meridan for any projection in ArcGIS. Is that what you are asking about?

As for reprojecting with your population grid, be careful. I know the LandScan pop grid warns against doing calculations on it in any projection other than WGS 84. This is becasue the grid is interpolated as it changes projections, skewing the data.

GIS Reference and Instruction Specialist, Stanford Geospatial Center.

www.mapbliss.com

### #7 frax Posted 04 December 2013 - 02:44 AM

frax

Hall of Fame

• 2,336 posts
• Gender:Male
• Location:Stockholm, Sweden
• Interests:music, hiking, friends, nature, photography, traveling. and maps!
• Sweden

I was just thinking today about how to do it using spherical coordinates and trigonometry, probably not too difficult if one has good enough math skills - but that would not be me!

Hugo Ahlenius
Nordpil - custom maps and GIS
http://nordpil.com/