Plotting a thematic map with a continuous color scale

General Usage
, ,
1

I would like to plot a thematic map (choropleth) using Plots.jl. The DataFrame df I use has 3 columns: geometry, value1 and value2. The column geometry is of element type Polygon (data was read using the Shapefile package). The columns value1 and value2 are of element type Int64 with values >= 0.

plot(df.geometry, color = :lightblue) creates a map with lightblue polygons without issues (and without colorbar to the side).

However, I would like the value2 column to color the polygons (shapes) based on a colormap/colorscheme like viridis with a corresponding colorbar to the side as legend.

Any suggestions how to do this?

plot(shp.geometry, palette = ColorSchemes.viridis) does plot a coloured map, however it is not clear on which variable this colours are based. I would like to specify that manually (value2 column) and have the colorbar added.

enter image description here

2

What happens if you add the plot argument: colorbar=true? Also, try argument: line_z = df.value2

3

Thanks for your reply. However, adding colorbar = true does not show a colorbar. And adding line_z = df.value2 does throw an error.

4

And fill_z = df.value2 ?

If these standard Plots.jl arguments do not work, I do not know nothing about Shapefile.jl objects and associated Plots’ recipes.

5

With zfill it does not plot anything (but gives no error message).

6

Here’s the full code for reproducibility and possible soltuions.

# Load modules
using CSV
using DataFrames
using Dates
using Shapefile
using ZipFile
using Plots
using ColorSchemes

# Create directorties
for dir in ["data", "downloads", "shapefiles"]
    path = joinpath(pwd(), dir)
    if !ispath(path)
        mkpath(path)
    end
end

# Download shapefiles
zip_url = "https://www.cbs.nl/-/media/cbs/dossiers/nederland-regionaal/wijk-en-buurtstatistieken/wijkbuurtkaart_2021_v1.zip" 
zip_loc = joinpath(pwd(), "downloads", split(zip_url, "/")[end])
if ~isfile(zip_loc)
	download(zip_url, zip_loc)
end


# Extract shape files
r = ZipFile.Reader(zip_loc)
for f in r.files
    file_name = split(f.name, "/")[end]
    if startswith(file_name, "gemeente")
        println("Extracting: $(file_name)")
        open(joinpath(pwd(), "shapefiles", file_name), "w") do io
            write(io, read(f))
        end
    end
end

# Read shapefiles
name_shapefile = "gemeente_2021_v1.shp"
path_shapefile = joinpath(pwd(), "shapefiles", name_shapefile)
table = Shapefile.Table(path_shapefile)
df = table |> DataFrame
row_filter = df.H2O .== "NEE"
df = df[row_filter, [:geometry]]

# Add columns value1 and value2
df.value1 = rand(1:100, nrow(df))
df.value2 = rand(1:5000, nrow(df))

# Plot map
plot(df.geometry, palette = ColorSchemes.viridis, axis=false, ticks=false, size=(500, 600))
7

I think you need to pass to the plot argument fill_z a transposed vector of values:
fill_z = permutedims(df.value1)

Another possibility, would be to define a customized vector of colors colorarray and pass it as a fill color argument:
fc = permutedims(colorarray)

1 Like
8

Thanks! fill_z = permutedims(df.value1) works for me and does show a nice colorbar :smiley:

2 Likes
9

fill_z = permutedims(df.value1) does work, however, I does not seem to use the specified colorscheme. Any suggestion on how to fix that?

10

I think you can use the plot argument color together with the palette of your choice. For example: c=palette(:viridis) or c=palette(:inferno)

2 Likes