极速阅览

📩 Load the geoJson file

This blogpost aims at building an hexbin map of US counties. Each county will be represented as a hexagone. The county boundaries are available here. For your convenience I've uploaded the exact file used in this post on github.

The coordinates are stored at geoJson format, probably the most common file format for this kind of information. If you want to know more about this, visit the background map section of the gallery that gives a tour on how to build a map with python.

Now, let's load this geoJson file from the web with geopandas

# Import geopandas library import geopandas as gpd # Load file url = "https://raw.githubusercontent.com/holtzy/The-Python-Graph-Gallery/master/static/data/us_states_hexgrid.geojson.json" geoData = gpd.read_file(url)

That's it, we now have a geo dataframe containing all the counties as hexagones. Before starting to draw a map with it, you can have a look at it with:

geoData.head(2)

cartodb_id created_at updated_at label bees iso3166_2 google_name geometry 0 1

1219	2015-05-13T22:26:15+00:00	2015-10-05T16:00:16+00:00	Maine	60.5	ME	Maine (United States)	POLYGON ((-72.62574 55.31320, -69.90286 54.408...
1217	2015-05-13T22:24:35+00:00	2015-10-05T16:00:16+00:00	R.I.	47.8	RI	Rhode Island (United States)	POLYGON ((-72.62574 49.57439, -69.90286 48.544...

🙇 Basic hexbin map

GeoPlot is probably the best python library to build a map from a geo dataframe. It is pretty straightforward thanks to its polyplot() function.

# Import the geoplot library. import matplotlib.pyplot as plt # Set the plot size for this notebook: plt.rcParams["figure.figsize"]=13,13 # Draw a map with matplotlib geoData.plot(color="white", edgecolor='black', linewidth=.5); plt.axis('off');

🖊️ Add county names

Let's add county names at the centroid of each. Fortunately, centroids are already available from the geometry column of the geo dataframe. For instance, you can access it for the first county with df.head(1).geometry.centroid.

Let's add a centroid column to the original dataframe:

# add a "centroid" column with the centroid position of each county geoData['centroid'] = geoData['geometry'].apply(lambda x: x.centroid) # redraw the empty hexbin map: geoData.plot(color="white", edgecolor='black', linewidth=.5); plt.axis('off'); # for each county, annotate with the county name located at the centroid coordinates for idx, row in geoData.iterrows(): plt.annotate(s=row['iso3166_2'], xy=row['centroid'].coords[0], horizontalalignment='center', va='center')

Now let's plot both the county borders and their labels at centroids position.

🛠️ Load & merge numeric data

Let's transform the useless hexbin map above in a choropleth map, where each county is colored according to its mariage rate (# of mariage per 1000 inhabitants). The dataset is available here

# Read the mariage data stored on github: import pandas as pd mariageData = pd.read_csv("https://raw.githubusercontent.com/holtzy/The-Python-Graph-Gallery/master/static/data/State_mariage_rate.csv") # Add a new column to the geo dataframe that will be used for joining: geoData['state'] = geoData['google_name'].str.replace(' \(United States\)','') # Merge the mariage dataset with the geospatial information geoData = geoData.set_index('state').join(mariageData.set_index('state'))

It is now possible to create a first choropleth map as follow:

geoData.plot(column="y_2015", cmap="viridis");

✨ Final map

Let's add the final touch to make it look good. Title, annotation, better color scale and more:

# Initialize the figure fig, ax = plt.subplots(1, figsize=(13, 13)) # map counties with the right color: geoData.plot( ax=ax, column="y_2015", cmap="BuPu", norm=plt.Normalize(vmin=2, vmax=13), edgecolor='black', linewidth=.5 ); # Remove useless axis ax.axis('off'); # Add title, subtitle and author ax.annotate('Mariage rate in the US', xy=(10,440), xycoords='axes pixels', horizontalalignment='left', verticalalignment='top', fontsize=14, color='black') ax.annotate('Yes, people love to get married in Vegas', xy=(10,420), xycoords='axes pixels', horizontalalignment='left', verticalalignment='top', fontsize=11, color='#808080') ax.annotate('python-graph-gallery.com', xy=(600,0), xycoords='axes pixels', horizontalalignment='left', verticalalignment='top', fontsize=8, color='#808080') # for each county, annotate with the county name located at the centroid coordinates for idx, row in geoData.iterrows(): ax.annotate( s=row['iso3166_2'], xy=row['centroid'].coords[0], horizontalalignment='center', va='center', color="white" ) # Add a color bar sm = plt.cm.ScalarMappable(cmap='BuPu', norm=plt.Normalize(vmin=2, vmax=13)) fig.colorbar(sm, orientation="horizontal", aspect=50, fraction=0.005, pad=0 );

Note: there are still some stuff that could be improved here but it's getting late 😴

Hexbin map from geoJson file with Python

📩 Load the geoJson file

🙇 Basic hexbin map

🖊️ Add county names

🛠️ Load & merge numeric data

✨ Final map

🚨 Grab the Data To Viz poster!