Python - Folium Choropleth Map - colors incorrect

user3121518 picture user3121518 · Oct 21, 2018 · Viewed 9k times · Source

My problem is that suburbs are not displaying the correct color on the Folium map. For example, Dandenong and Frankston should be shaded with the darkest color as they have the highest count in the dataframe, but they are shaded with a lighter color.

The dataframe is missing some suburbs. Those suburbs are being colored with the darkest color.

Another problem is the csv has all suburbs in UPPERCASE but the geojson has a mixture of cases such as "Frankston", "St Kilda", or "McKinnon". It would be helpful if the choropleth code didn't care about case. I can change the text in the dataframe to make "FRANKSTON", "Frankston", and "ST KILDA", "St Kilda", but "MCKINNON" to "McKinnon" is proving a bit trickier.

Create Dataframe

import csv 
import pandas as pd
csv_path='Data_tables_Criminal_Incidents_Visualisation_year_ending_June_2018.csv'
df=pd.read_csv(csv_path)

with open(csv_path, 'r') as csvfile: 
    # creating a csv reader object 
    csvreader = csv.reader(csvfile) 
    # create a list of headings from the first row of the csv file
    headings = next(csvreader)

# create a dictionary, where keys are Suburb/Town Name and values are number of occurences
# index 2 of the headings list are the suburbs
neighborhood_dict = df[headings[2]].value_counts().to_dict()

# make first letter uppercase eg St Kilda
neighborhood_dict = dict((k.title(), v) for k, v in neighborhood_dict.items())


# make neighborhood_list from neighborhood_dict
neighborhood_list=[]
for key, value in neighborhood_dict.items():
    temp = [key,value]
    neighborhood_list.append(temp)

# make dataframe from neighborhood_list
df = pd.DataFrame(neighborhood_list, columns=['Suburb','Count'])

print(df.to_string()) 

Create Map

import folium

world_map = folium.Map(
        location=[-38.292102, 144.727880],
        zoom_start=6,
        tiles='openstreetmap'
        )

world_map.choropleth(
        geo_data='vic.geojson',
        data=df,
        columns=['Suburb','Count'],
        key_on='feature.properties.Suburb_Name',
        fill_color='YlOrRd',
        fill_opacity=0.7,
        line_opacity=0.2,
        legend_name='Crime Rate in Victoria'
        )

world_map.save('index.html')

Dataframe Image

Legend

Map Image

Answer

user3121518 picture user3121518 · Oct 25, 2018

I got it all figured out. Missing values are coloured grey, and the legend is customized with intervals of my choice. Cleaning up the geojson, removing trailing white space, and making all suburb names UPPERCASE solved a lot of problems.

Files are here

Demo image

Create Dictionary

import pandas as pd
import csv 

csv_path='Data_tables_Criminal_Incidents_Visualisation_year_ending_June_2018.csv'
df=pd.read_csv(csv_path)

# sum the number of incidents recorded for each suburb
df=df.groupby(['Suburb/Town Name'])['Incidents Recorded'].agg(
    # make the numbers numeric otherwise it just concatenates strings
    lambda x: pd.to_numeric(x, errors='coerce').sum()
)

# create a dictionary, where keys are Suburb/Town Name and values are number of incidents
suburb_dict = df.to_dict()

Style Function

def style_function(feature):
    suburb = suburb_dict.get(feature['properties']['Suburb_Name'])
    return {
        'fillColor': '#gray' if suburb is None else colormap(suburb),
        'fillOpacity': 0.6,
        #borders
        'weight': 0.2,
    }

Folium Map

import folium

world_map = folium.Map(
        location=[-38.292102, 144.727880],
        zoom_start=6,
        tiles='openstreetmap'
        )

folium.GeoJson(
    data = 'vic_for_crime_2018.geojson',
    style_function = style_function    
).add_to(world_map)

Colormap

import branca

colormap = branca.colormap.linear.YlOrRd_09.scale(0, 8500)
colormap = colormap.to_step(index=[0, 1000, 3000, 5000, 8500])
colormap.caption = 'Incidents of Crime in Victoria (year ending June 2018)'
colormap.add_to(world_map)

world_map.save('vic_final.html')