11 replies to this topic

[A. Fenix]

Hello fellow mappers,

I am working on a cartogram map of the worlds caloric intake but keep getting unsavory results (pun intended). This is my first jaunt down cartogram lane, and have tried a couple of different scripts but neither are giving me the results I would have expected. I have tried Eric Wolf's Cartogram Creator script, as well as Tom Gross's Cartogram Geoprocessing Tool Version 2. <quick pause to give props to both of these script writers...>, but neither tools output is giving me the results I would have expected. For example, the two countries that have the highest daily caloric intake are the U.S. and Austria (both are over 3500..), and yet the U.S. is shrinking in my results, while Europe is inflating. The latter makes sense since many of the top caloric intake countries are in Europe, but again, the U.S. is #1 (and Canada #13) and yet N. America is shrinking (see attached). After toying with this for awhile I began to wonder if the initial area of the countries may be effecting the results in some way, but since I'm simply trying to do value-by-area map I am surprised that N. America is shrinking... any thoughts from fellow cartogram guru's out there?

btw, I'm using ArcGIS 9.3.1 SP2, ESRI Country boundaries and FAO data as inputs...

I hope all are well!

Analisa

Attached Files

•  bad_cartogram.jpg   306.48K   71 downloads

[dsl]

My initial reaction is to say that it has to do with the size of the country (Austria being significantly smaller than North America). You might try a different projection that more accurately represents the size/shape of the country. I'm not sure which you are already using. You may also want to make sure that your polygons are properly dissolved so that there is one per country? I don't know if the scripts handle multipolygons though.

Hope that helps,
David

[A. Fenix]

Thanks for the feedback David. Unfortunately, I have tried both of those suggestions, with the output remaining more or less the same. I have tried a number of different projections, and even completely reworked the input country data so that there was only one polygon per country (which of course would ruin the output results anyway, since countries like Canada, Indonesia, etc would completely loose their shape) just to see if this would help, but to no avail. I keep looking longingly at the fantastic worldmapper site (http://www.worldmapper.org), and am just baffled but what misstep I have inadvertently taken...

ce la vie of a mapper...

analisa

[dsl]

Maybe try normalizing the data in some way, e.g. caloric intake per thousand persons....I'm not sure what a standard measure would be for that particular outcome.

Edit:

I looked through the world mapper site, you might want to try and create a variable that relates all the countries together, like the USA's proportion of global caloric intake...

[frax]

I would start with a projection that preserves shapes fairly well, like Miller Cylindrical, and then you might want to move some countries so that they don't get stretch out in a funny way (no shared boundaries). You want to preserve the shape of the countries as much as possible.

See this one that I made a while ago:
http://nordpil.com/g...untriesregions/

[A. Fenix]

Thanks everyone for the great feedback. Each of the suggestions played a roll in getting my cartogram map to work. The issues were many, but the key that I took away from this cartogram adventure was knowing what types of data are appropriate, and/or how to prepare your data to make it appropriate for cartogram air time. I was fortunate to have an email exchange with Benjamin D. Hennig, one of the primary cartogrammers for world mapper who echoed DSL about how I could remedy my cartogramming mishap: cartograms require absolute, not relative data. As Ben put it, "If you think of a cartogram as a geographic version of a pie chart it should become clearer what kind of data you need to create a cartogram-style map." Therefore, the primary issue was that I had relative data since it is the average Caloric Intake per person each day for each country. In order to make it absolute, and to keep in line with the theme of the map, I made it absolute by multiplying the average with the population of each country, thereby creating a second theme; the total amount of calories consumed by each country. Since this was not the whole story I was attempting to tell with the data, I used chloropleth shading to show the daily caloric intake by country. There were a number of other things that had to be done to pertaining to finding the appropriate projection, etc. in order to find the precarious balance between maintaining the countries shape enough that it can be recognized by the map ready, yet being as true to the data as possible. After working on this cartogram, it reminded me a great deal of the careful thought that must be put into which projection one is going to use for analysis and mapping purposes. There's always the balancing act of area, distance and shape. No projection is completely "true" to reality. However, I felt that working in the cartogram realm was a bit more of a black box. I tweaked everything I could in the program, ran countless iterations, investigated and reinvestigated the results, and compared endlessly. In the end, I am happy with the results, but have a new appreciation for the strengths, as well as weaknesses of cartograms.

Special thanks again to Ben, DSL and the programmers who have created so many fantastic cartogram creators (in the end I found ScapeToad to be the best suited for my needs).


You can see the resultant cartogram map attached.

[A. Fenix]

it appears there's been some changes to how the cartotalkers upload...so hopefully this will work. Attachment as promised:  CalorieCartogram_afenix0829.jpg   674.85K   40 downloads

[A. Fenix]

the above is NOT at all the correct color palette. Lately I've been having some issues with export and screen resolution. hrmph. Needless to say, here is a closer approximation of the final print colors:

 CalorieCartogram_afenix0829.png   255.05K   54 downloads

[DaveB]

Nice work. I like how the grid lines help emphasize which countries are pinched in and which bulge out.

The title should say "... Each Country's Population". The possessive, not the plural. Or maybe it can be worded better. I'm not quite sure what it means. From looking at the map I would say as a country India consumes more total calories than any other country.
Although I like the grid lines they do get a bit in the way when trying to read labels in the ocean and the legend text.
The colors/legend represent the per capita intake? It's not specified anywhere.
Country labels seem a bit random, both in which ones are labeled (I suppose there just there to help readers figure out other countries, too), and in their placement. Some straight, some curved.

[frax]

So what is it that you want to communicate? - what is the core message that you want to convey to the target audience, and how should they interpret the map?

The thing is - if you start to ask questions like these, you will soon realize that cartograms works best as a gimmick, and are very heard to interpret...

Eurasia, Scandinavia and Canada often gets very distorted (esp Russia). The regional differences between Northern Africa, Sub-Saharan Africa and South Africa as an outlier often results in quite a squished shape.

[frax]

Not that gimmicks are all that bad, of course - they have a place too!

[A. Fenix]

Thanks for the suggestion. This map will be embedded within an internal publication describing regional resilience and vulnerability. The map is an attempt to inspire readers to think about the complexity of food issues around the globe (such as variation in consumption patterns as this map is working to explore). As for the title, I agree! It's not meant to be in the publication, but was put there as shorthand instructions for the senior writer/editor of the publication. The map will have a caption which will be explaining the all elements (such as per capita intake). Yes, cartograms can be thought of as "gimmicky" I suppose. However, I see them more as yet another palette from which to tell a story. As a first attempt (and second draft waiting for final touch ups to fit in with publication) I'm happy with the results, but again, having experienced the creation process I will turn a more thoughtful and cautious gaze at all cartograms I see, and will make them a discussion piece when lecturing on cartography in the future.

Thanks for the fine thoughts and suggestions Dave and Frax!