Choropleth geojson fitbounds with wrapping geojson

đź“Š Plotly Python
1

I’m having an issue with wrapping and geojson.

Here’s an example based on Choropleth map using GeoJSON (https://plotly.com/python/choropleth-maps/)

from urllib.request import urlopen
import json
with urlopen('https://raw.githubusercontent.com/plotly/datasets/master/geojson-counties-fips.json') as response:
        counties = json.load(response)

import pandas as pd
df = pd.read_csv("https://raw.githubusercontent.com/plotly/datasets/master/fips-unemp-16.csv",
                 dtype={"fips": str})

import plotly.express as px

df = df[df.fips.str.startswith("01")]

fig = px.choropleth(df, geojson=counties, locations='fips', color='unemp',
                    color_continuous_scale="Viridis",
                    range_color=(0, 12),
                    labels={'unemp':'unemployment rate'}
)
fig.update_geos(fitbounds="locations", visible=False)

import dash
import dash_core_components as dcc
import dash_html_components as html

app = dash.Dash()
app.layout = html.Div([
        dcc.Graph(figure=fig)
    ])

app.run_server(debug=True, use_reloader=False)

From the code you can see I filter down to the Alabama data (FIPS state code 01) and use fitbounds=“locations” to center and zoom on the provided data. And you can see in the picture below that it works great.

alabma
alabma700Ă—450 26.7 KB

Now, let’s try with Alaska (FIPS 02), as in

df = df[df.fips.str.startswith("02")]

alaska
alaska700Ă—450 8.34 KB

Gross! I’ve tried all manner of update_geos with center, projection_rotation, *axis_range, etc. but the fitbounds overrides them all, and without it I have to manually calculate and set everything (assuming I can replicate all the zooming too).

Is there any way to have fitbounds locations understand that a rotation before bounding is better than wrapping? Or, let me provide a center or rotation prior to calculating bounds?

2

Using the following special case code for Alaska gets me close to what I want. However, I think any geojson shapes that wrap will have this issue.

if alaska:
        fig.update_geos(
              lonaxis_range=[20, 380],
              projection_scale=6,
              center=dict(lat=61),
              visible=False)
else:
        fig.update_geos(fitbounds="locations", visible=False)

alaska2
alaska2700Ă—450 25.3 KB

Any advice on a better way to do this so that a special case isn’t required? Or could the fitbounds location code be updated to handle this?

3

I had the same problem. Thanks for posting this. It would be great to get an update on how to fix this. Additionally, I’m struggling with keeping the aspect ratios 1:1. My Alaska looks stretched like yours. I have tried everything to try and get it to maintain the aspect ratio. Any guidance on that would be great.

4

Unfortunately, I have no update. Mapping a 3D object (the globe) to a 2d surface is just hard. Maintaining shape, distance, and relative size all at once is impossible and you have to be will to bend a little on one of them. This is true of all map projections and there are some great books about cartographic projections.

More importantly you’re really asking about better fixes to the problem I outlined above. For my use case the workaround I posted was “good enough.” So I stopped there and haven’t revisited. If you come up with something better I’d love to see it. Or if someone else from the community wants to chime in, that’d be great too.

Anyone?