Size of bubbles in plotly.express.scatter_mapbox

Question

I am trying to create a map with plotly.express that displays values are bubbles (circles).

The values currently range from 16000 to 21500. I have got everything up and running, the bubbles are displayed in different colors, however, they are more or less all of the same size.

What I would like to have is the smallest value to be displayed with a small bubble and the largest value with a large bubble and the other values in between.

This is how my data frame looks like:

                 country       average       long        lat
0        Baden-Württemberg  19166.381092   9.179330  48.781956
1                   Bayern  18786.556728  11.572199  48.137859
2                   Berlin  21463.044514  13.387224  52.533707
3              Brandenburg  19622.567766  13.070526  52.405476
4                   Bremen  16197.013903   8.805129  53.081386
5                  Hamburg  18426.436184  10.001104  53.554158

and this is the way how I display it:

fig = px.scatter_mapbox(all_data, lat="lat", lon="long", hover_name="country", hover_data=["country", "average"], 
                        color="average",
                        size="average", color_continuous_scale=px.colors.sequential.matter, size_max=20,
                        zoom=5, height=1000, mapbox_style="open-street-map")
fig.update_layout(margin={"r":0,"t":0,"l":0,"b":0})
fig.show()

And this is how it looks like: open street map with bubbles

How can I influence the size of the bubbles so that a smaller value has a small diameter and and larger value has a large diameter?

I tried to play with the size_max-value, but all the bubbles will still have the same size, just all of them larger or smaller.

Answer 1

I have found out, that the size-Parameter can take a list of values that reflect the scale. This will not influence the scale that is drawn right of the map.

So I did this:

# doing a little bit of math here to calculate a scale to reflect the difference between
# the minimum and the maximum of the average prices (could probably be done much more elegant,
# but this does the job)
# 
# first, calculate a ratio between max and min and divide it to have 16 steps

all_data_diffq = (all_data["mean"].max() - all_data["mean"].min()) / 16

# calculate the scale value by subtracting the minium value from the average price, divide 
# that by the ratio which will give the scale a value between 0...15 and add 1 to it so that
# the scale values start at 1 (to be visible on the map)
# add the according scale to each row
# the scale column will then be used for size=... parameter in the scatter_mapbox call below

all_data["scale"] = (all_data["mean"] - all_data["mean"].min()) / all_data_diffq + 1

My dataframe looks like this now:

                   country          mean       long        lat      scale
0        Baden-Württemberg  19166.381092   9.179330  48.781956  10.021952
1                   Bayern  18786.556728  11.572199  48.137859   8.867916
2                   Berlin  21463.044514  13.387224  52.533707  17.000000
3              Brandenburg  19622.567766  13.070526  52.405476  11.408003
4                   Bremen  16197.013903   8.805129  53.081386   1.000000
5                  Hamburg  18426.436184  10.001104  53.554158   7.773747

And the call of scatter_mapbox() now uses the "scale" column for the size-parameter:

fig = px.scatter_mapbox(all_data, lat="lat", lon="long", hover_name="country", hover_data=["country", "mean"], 
                        color="mean",
                        size=all_data["scale"], color_continuous_scale=px.colors.sequential.Rainbow,
                        size_max=50, zoom=5, height=1000, mapbox_style="open-street-map")
fig.update_layout(margin={"r":0,"t":0,"l":0,"b":0})
fig.show()

Now the result looks better:

applied scale the to the size-parameter

Answer 2

Your code is correct. The size was not noticeable due to the similarity of the data subject to the size. I deliberately modified Bremen's data to a larger size to draw the graph. The other fix is to change the 'size_max'. I also changed the color of the bubble to a color that does not blend with the map.

import pandas as pd
import numpy as np
import io

data = '''
 country average long lat
0 Baden-Württemberg  19166.381092 9.179330 48.781956
1 Bayern 18786.556728 11.572199 48.137859
2 Berlin 21463.044514 13.387224 52.533707
3 Brandenburg 19622.567766 13.070526 52.405476
4 Bremen 46197.013903 8.805129 53.081386 # average value update
5 Hamburg 18426.436184 10.001104 53.554158
'''

all_data = pd.read_csv(io.StringIO(data), sep='\s+')

import plotly.express as px

fig = px.scatter_mapbox(all_data, lat="lat", lon="long", hover_name="country", hover_data=["country", "average"], 
                        color="average",
                        size="average", color_continuous_scale=px.colors.sequential.Rainbow, size_max=40,
                        zoom=5, height=1000, mapbox_style="open-street-map")
fig.update_layout(margin={"r":0,"t":0,"l":0,"b":0})

fig.show()