Weird Scattermapbox lines behaviour

Dash Python
1

Screenshot 2023-03-23 at 13.48.55
Screenshot 2023-03-23 at 13.48.551051×331 37.6 KB

See attached screenshot. How can I fix that weird behaviour? All coords I submit are “in the water” and at 180deg lon there’s a cutoff (they are not connected). It works perfectly fine for every set of coords which “does not cross the pacific”.

Here’s the code:

def update_map_fig(datastore):
        waypoints=pd.DataFrame(datastore)
        map_fig_updated = go.Figure(go.Scattermapbox(
        mode="lines",
        lon=waypoints.lon.values, lat=waypoints.lat.values))
        map_fig_updated.update_layout( 
        mapbox=dict(accesstoken="mytoken"), margin={"r": 0, "t": 0, "l": 0, "b": 0},showlegend=False)
        return map_fig_updated
2

If you change the mode to ‘markers’ are the markers plotted in the correct spots?

2 Likes
3

good question! Yes, the markers are on the correct spot:

Screenshot 2023-03-23 at 15.36.01
Screenshot 2023-03-23 at 15.36.011037×331 85.3 KB

4

I think the lines could be an issue with how the map works.

What happens when you zoom in with the lines?

1 Like
5

Screenshot 2023-03-23 at 15.41.36

6

Very strange…

Might be an issue with the API for mapbox…

7

In a scatter plot the order of the points defines the lines to be drawn.

I hade something like this in the past:

8

that cutoff happens right at 180 deg long (and is obviously not visible until fully/max zoomed in)

9

Since this problem is exclusive to lines that go “around the world”, meaning lines that cross 180 deg lon (and I’m pretty sure the order is correct too) I don’t think that’s it

10

Yeah, you might be right, just wanted to mention it :wink:

11

So I have found out that this is in fact a feature, not a bug. Crossing the antimeridian - any tips on how to proceed?

12

Here’s the solution for future reference:

        #waypoints is a pandas df with columns lon and lat in degrees ranging from -180 to 180
        waypoints=pd.DataFrame(waypoints)
        diffs=np.diff(lons:=waypoints.lon.values)
        #the number 180 depends on your "step size", may need customization
        crossings_plusminus=np.where(diffs<=-180)[0]
        crossing_minusplus=np.where(diffs>180)[0]
        for plusmin_crossing in crossings_plusminus:
            lons[plusmin_crossing+1:]+=360
        for minusplus_crossing in crossing_minusplus:
            lons[minusplus_crossing+1:]-=360
        fig = px.line_mapbox(lat=waypoints.lat, lon=lons)
1 Like
13

Thanks for sharing the solution.