Folium and Geopandas: FeatureGroup for categorial data

Folium is a Python library for creating interactive maps. The plotted markers or polygons can be assigned to individual layers (using folium.FeatureGroup()), which can be switched on and off with a mouse click.

All the examples I found on the net explicitly define each individual layer with a command like:

 feature_group1 = FeatureGroup(name='Foo')

If I have categorical data in a Geopandas.GeoDataFrame, I certainly don’t want to create a layer for each category by hand. Fortunately, we can automate this by simply appending each initialised FeatureGroup to a list. Here’s an example with volcanoes.

A quick look at the data:

 volcanoes.info()


RangeIndex: 1233 entries, 0 to 1232
Data columns (total 13 columns):
 #   Column             Non-Null Count  Dtype   
---  ------             --------------  -----   
 0   number             1233 non-null   int64   
 1   name               1233 non-null   object  
 2   country            1233 non-null   object  
 3   vtype              1233 non-null   object  
 4   evidence           1233 non-null   object  
 5   last_eruption      1233 non-null   object  
 6   region             1233 non-null   object  
 7   subregion          1233 non-null   object  
 8   elevation          1233 non-null   int64   
 9   rocks              1198 non-null   object  
 10  tectonic_setting   1230 non-null   object  
 11  last_eruption_int  787 non-null    float64 
 12  geometry           1233 non-null   geometry
dtypes: float64(1), geometry(1), int64(2), object(9)
memory usage: 125.4+ KB

I want a map grouped by rock type.

grouped = volcanoes.groupby('rocks')

Since Seaborn’s colour palettes are particularly nice, I use them to create a list of colours in hex format (as used by html). The number of colours needed is len(grouped).

import seaborn as sns
pal = sns.color_palette("husl", len(grouped)).as_hex() 

In this example I do not want the base map (the tile layer) to be displayed in the LayerControl as well. Therefore I initialise the map without tiles and add a TileLayer with control=False:

m = folium.Map(tiles=None)
folium.TileLayer('cartodbpositron', control=False).add_to(m)

Now we iterate through the grouped GeoDataFrame. For each category a FeatureGroup is created, which is appended to the list f_groups. Then markers for corresponding volcanoes are added to the last initialised FeatureGroup.

f_groups = []

for group_name, group_data in grouped:
    f_groups.append(folium.FeatureGroup(group_name))
    color = pal.pop()
    
    for i in range(0,len(group_data)):

        # html for popup of markers
        html=f"""
            <h2>  {group_data.iloc[i]['name']} </h2>
            <small>
            <p> Country: {group_data.iloc[i]['country']}  <br/>
            Elevation: {group_data.iloc[i]['elevation']}  <br/>
            Last Eruption: 
            {group_data.iloc[i]['last_eruption']}  <br/>
            Rocks: {group_data.iloc[i]['rocks']}  <br/>
            Tectonic Setting: 
            {group_data.iloc[i]['tectonic_setting']}  </p>
            </small>
            """

        iframe = folium.IFrame(html=html, width=300, height=200)
        popup = folium.Popup(iframe, max_width=650)

        # Add markers to last FeatureGroup    
        folium.CircleMarker(
            location=[group_data.iloc[i].geometry.y, 
                      group_data.iloc[i].geometry.x],
            radius=5,
            popup=popup,
            tooltip=group_data.iloc[i]['name'],
            fill_color=color,
            stroke = False, 
            fill_opacity = 1,
        ).add_to(f_groups[-1])

    # Add last featureGroup to Map
    f_groups[-1].add_to(m)

folium.LayerControl().add_to(m)

m