Poltly Express `scatter_mapbox` with `animation_frame` and `color`

Question

If have data with columns : 'e_lat', 'e_lon', 'Année' and 'Expéditeur'. I'm trying to make a scatter_mapbox, with frames animated by 'Année", and color based on 'Expéditeur' (which is categorical, the names of the persons).

# I created a row df['_color'] = df['Expéditeur'] because in the end I want to change the way _color is defined.

fig = px.scatter_mapbox(df,
                        lat='e_lat',
                        lon='e_lon',
                        hover_name='Ville Expéditeur',
                        hover_data=['Expéditeur', 'Destinataire', 'Ville Expéditeur', 'Année'],
                        animation_frame='Année',
                        mapbox_style='carto-positron',
                        zoom=4,
                        color='_color',
                        )
fig.show()

The trouble is I get something like this : (only the points that have the first color are showing)

If I comment out color='_color',, I get : If I comment out animation_frame='Année'',, I get :

I did a lot of research but it always seems to me that people are doing exactly the same and have different results ([this tutorial for example](https://towardsdatascience.com/how-to-animate-scatterplots-on-mapbox-using-plotly-express-3bb49fe6a5d, this stackoverflow issue, ...)

My plotly lib is up-to-date.

I tried a lot of different things with `animation group, color_discrete_map, color_discrete_sequence etc. but without any success).

What am I doing wrong here ? Thx a lot.

---

[EDIT] In fact it's exaclty the same problem than here (not resolved). This seems to be because of a documented bug of Plotly. "If all colors are used for the first frame, the problem disappears."

I solved it by adding "fake data" (outside of the scope of the map) : one for each color, in the first frame of the animation.

new_rows = []
for color in df['_color'].unique():
    new_row = {'_color': color, 'Année': df['Année'].min(), 'lat': 0, 'lon': 0}
    new_rows.append(new_row)
df = df.append(new_rows, ignore_index=True)

Hope this will help someone.

Answer 1

In fact it's exaclty the same problem than here (not resolved). This seems to be because of a documented bug of Plotly. "If all colors are used for the first frame, the problem disappears."

I solved it by adding "fake data" (outside of the scope of the map) : one for each color, in the first frame of the animation.

new_rows = []
for color in df['_color'].unique():
    new_row = {'_color': color, 'Année': df['Année'].min(), 'lat': 0, 'lon': 0}
    new_rows.append(new_row)
df = df.append(new_rows, ignore_index=True)

Hope this will help someone.