Guide to processing maps from GADM to make cartograms

This is the process for preparing map data to generate cartograms.

Table Of Contents

• Step 1: Download map from GADM
• Step 2: Process map in Mapshaper
  • Step 2.1: Projection
  • Step 2.2: Simplification
  • Step 2.3: Clipping
  • Step 2.4: Export
• Step 3: Generate CSV and JSON file
  • Errors?
• Optional: What if my map is NOT from GADM?
• Optional: What if I need to remove regions?

Step 1: Download map from GADM

First, visit https://gadm.org/download_country_v3.html and select the country. Click on "Shapefile", which will download a zip file.

Step 2: Process map in Mapshaper

Then, visit https://mapshaper.org/ and import the zip file. Click on the correct layer, which should be gadm36_[country]_1, as seen in the image below:

Step 2.1: Projection

You will need to know the desired projection for this map. Then, click on console (upper right corner) and type the following command to project the map, replacing [spatial_reference] with your chosen spatial reference system:

$ -proj +init=EPSG:[spatial_reference]

Check to make sure that the map is projected correctly.

Step 2.2: Simplification

Type the following command to find out the total number of vertices.

$ -simplify 100% stats

The unique coordinate locations tells you the number of vertices. Use the following equation to calculate the percentage required.

[percentage] = 100 * [desired no. of vertices] / [unique coordinate locations]

So, if you would like 50,000 verticies, then the percentage will be 100 * 50,000 / 573,076 ≈ 9

Then, type the following command to execute the simplification, replacing [percentage] with the required percentage:

$ -simplify [percentage]% stats

Step 2.3: Clipping

Now, click on export at the top right corner. Select ONLY the correct layer gadm36_[country]_1 and click on SVG file format.

Open the SVG file in Inkscape and check whether the bounding box is sufficiently tight. If yes, proceed to Step 2.4: Export. If not, continue with the following steps.

As we can see in the image above, there is unnecessary white space on the right side of the image. We would like to get rid of that.

In Mapshaper, type the following command:

$ info

You will see a bunch of information pop up. Scroll to the correct layer (make sure it is the right one!) and copy the bounds information.

Then, type the following command but don't press enter yet.

$ -clip remove-slivers bbox=[paste bounds here]

For example, for Japan, we would get this:

$ -clip remove-slivers bbox=-819424.5537347307,-198081.9783434747,2375994.3457566854,2224098.3370094863

Before pressing enter, you need to edit the command. -clip will remove everything that is outside the specified bbox. The bbox is in the format xmin, ymin, xmax, ymax. Since the extra space is in the right side of the map, we need to reduce xmax. We can check what value to put in by placing our cursor over the map. Place your cursor to the rightmost edge which you wish to retain (Refer to the cursor in the image below). Read the x value, round it to give a bit of allowance (round up for xmax and ymax, round down for xmin and ymin), and put it in the command, replacing the old value.

For example, for Japan, we could change it to this:

$ -clip remove-slivers bbox=-819424.5537347307,-198081.9783434747,1210000,2224098.3370094863

You can now press enter. Download the SVG file and check the bounding box again. Repeat as necessary. In this image below, we could remove the islands in the bottom left if we wanted to. We could also keep it in. Once it is satisfactory, proceed to the next step.

Step 2.4: Export

In Mapshaper, export the map using the following command, replacing [country_name] accordingly.

$ -o format=geojson bbox precision=0.1 [country_name].json

Step 3: Generate CSV and JSON file

Open geojson2csv_cartogram-web.r in RStudio. Make sure that the GeoJSON file which you just downloaded is in the same folder. Rename file_name accordingly. Then, click on source. In the console, it should say "All done.".

You will be able to find [country]_data.csv in your folder. Here, you should check whether the region names are accurate. Open the csv file using a text editor or spreadsheet program, and make sure that you save in the csv format subsequently. Fill in the Region.Data column with something other than population. At this stage, you should also check whether the region names and abbreviations are correct by comparing with an official source. Editing the names/abbreviations in the csv file is sufficient to correct any typos.

After saving, you now have your [country]_processedmap.json and [country]_data.csv files ready to use to generate cartograms!

Please proceed to the Add Map Guide to continue the process.

Errors?

If you do not have the rgdal package installed, install it using install.packages("rgdal").

If you encounter "Error in ogrListLayers(dsn = dsn) : Cannot open data source", check that the file_name is correct, and that the json file and R script are in the same folder. Also, make sure that you click Session --> Set Working Directory --> To Source File Location.

Optional: What if my map is NOT from GADM?

If your map is not from GADM, you will see the following error in RStudio:

To fix this, look through the properties and values and figure out which property contains the region name and region abbreviation.

Then, edit "nameproperty" and "abbrproperty" accordingly, as seen below.

Then, re-run the program (click on source).

Optional: What if I need to remove regions?

If you need to remove regions, like a small island territory that has a population of 4 people, then perform the following step before Step 2.1: Projection.

You should know the name of the region which you would like to remove. Then, in mapshaper's console, type in the following command, replacing [region_name] accordingly:

$ filter 'NAME_1 != "[region_name]"'

If successful, you should see the following message: "[filter] Retained XX of XX features"