Contents

Employment rates in Finland

Goal: plot an interactive map of employment rates across Finnish regions.

Required modules:

  • Folium for plotting interactive maps based on leaflet.js

  • Pandas for handling tabular data

  • Geopandas for handling spatial data

# Import required modules:
import folium
import pandas as pd
import geopandas as gpd
import matplotlib.pyplot as plt
%matplotlib inline

Employment rate data

Employment rate refers to “the proportion of the employed among persons aged 15 to 64”. Data from Statistics Finland (saved from .xml-file to csv in Excel..).

# Read in data
data = pd.read_csv("data/seutukunta_tyollisyys_2013.csv", sep=",")
data.head()
seutukunta seutukunta_nimi tyollisyys
0 SK011 Helsingin seutukunta 73.0
1 SK014 Raaseporin seutukunta 70.3
2 SK015 Porvoon seutukunta 74.3
3 SK016 Loviisan seutukunta 71.5
4 SK021 Åboland-Turunmaan seutukunta 72.9

Sub-regional units

The spatial data for the sub-regional units (Seutukunnat in Finnish) can be retrieved from the Statistics Finland Web Feature Service http://geo.stat.fi/geoserver/tilastointialueet/wfs

# A layer saved to GeoJson in QGIS..
#geodata = gpd.read_file('Seutukunnat_2018.geojson')

# Get features directly from the wfs
url = "http://geo.stat.fi/geoserver/tilastointialueet/wfs?request=GetFeature&typename=tilastointialueet:seutukunta1000k_2018&outputformat=JSON"
geodata = gpd.read_file(url)

geodata.head()
id vuosi seutukunta nimi namn name geometry
0 seutukunta1000k_2018.1 2018 011 Helsinki Helsingfors Helsinki MULTIPOLYGON (((409963.522 6681658.341, 409969...
1 seutukunta1000k_2018.2 2018 014 Raasepori Raseborg Raasepori MULTIPOLYGON (((306616.919 6665438.489, 306668...
2 seutukunta1000k_2018.3 2018 015 Porvoo Borgå Porvoo MULTIPOLYGON (((427108.141 6694151.025, 427175...
3 seutukunta1000k_2018.4 2018 016 Loviisa Lovisa Loviisa MULTIPOLYGON (((444038.768 6703649.355, 444155...
4 seutukunta1000k_2018.5 2018 021 Åboland-Turunmaa Åboland-Turunmaa Åboland-Turunmaa MULTIPOLYGON (((190999.717 6715878.622, 191021...

Join attributes and geometries

We can join the attribute layer and spatial layer based on the region code (stored in column ‘seutukunta’). The region codes in the csv contain additional letters “SK” which we need to remove before the join:

data["seutukunta"] = data["seutukunta"].apply(lambda x: x[2:])
data["seutukunta"].head()

0    011
1    014
2    015
3    016
4    021
Name: seutukunta, dtype: object

Now we can join the data based on the “seutukunta” -column. Let’s also check that we have a matching number of records before and after the join:

#print info
print("Count of original attributes:", len(data))
print("Count of original geometries:", len(geodata))

# Merge data
geodata = geodata.merge(data, on = "seutukunta")

#Print info
print("Count after the join:", len(geodata))

geodata.head()

Count of original attributes: 70
Count of original geometries: 70
Count after the join: 70
id vuosi seutukunta nimi namn name geometry seutukunta_nimi tyollisyys
0 seutukunta1000k_2018.1 2018 011 Helsinki Helsingfors Helsinki MULTIPOLYGON (((409963.522 6681658.341, 409969... Helsingin seutukunta 73.0
1 seutukunta1000k_2018.2 2018 014 Raasepori Raseborg Raasepori MULTIPOLYGON (((306616.919 6665438.489, 306668... Raaseporin seutukunta 70.3
2 seutukunta1000k_2018.3 2018 015 Porvoo Borgå Porvoo MULTIPOLYGON (((427108.141 6694151.025, 427175... Porvoon seutukunta 74.3
3 seutukunta1000k_2018.4 2018 016 Loviisa Lovisa Loviisa MULTIPOLYGON (((444038.768 6703649.355, 444155... Loviisan seutukunta 71.5
4 seutukunta1000k_2018.5 2018 021 Åboland-Turunmaa Åboland-Turunmaa Åboland-Turunmaa MULTIPOLYGON (((190999.717 6715878.622, 191021... Åboland-Turunmaan seutukunta 72.9 
## Create a static map

Now we have a spatial layer with the employment rate information (in column “tyollisuus”). Let’s create a simple plot based on this data:

# Define which variable to plot
geodata.plot(column="tyollisyys")

<matplotlib.axes._subplots.AxesSubplot at 0x20e085240c8>
../../_images/Employment_in_Finland_14_1.png

Adjusting the figure, we need to import matplotlib pyplot

# Adjust figure size
fig, ax = plt.subplots(1, figsize=(10, 8))

# Adjust colors and add a legend
geodata.plot(ax = ax, column="tyollisyys", scheme="quantiles", cmap="Reds", legend=True)

<matplotlib.axes._subplots.AxesSubplot at 0x20e085d4c08>
../../_images/Employment_in_Finland_16_1.png

Create an interactive map

Next, we’ll plot an interactive map based on the same data, and usign the folium library, which enables us to create maps based on the JavaScript library leaflet.js.

# Create a Geo-id which is needed by the Folium (it needs to have a unique identifier for each row)
geodata['geoid'] = geodata.index.astype(str)

# Create a Map instance
m = folium.Map(location=[60.25, 24.8], tiles = 'cartodbpositron', zoom_start=8, control_scale=True)

folium.Choropleth(geo_data = geodata, 
                  data = geodata, 
                  columns=['geoid','tyollisyys'], 
                  key_on='feature.id', 
                  fill_color='RdYlBu', 
                  line_color='white',
                  line_weight=0,
                  legend_name= 'Employment rate in Finland').add_to(m)
m

We can also plot “tooltips” on the map, which show the values for each feature.

folium.features.GeoJson(geodata, name='Labels',
               style_function=lambda x: {'color':'transparent','fillColor':'transparent','weight':0},
                tooltip=folium.features.GeoJsonTooltip(fields=['tyollisyys'],
                                              aliases = ['Employment rate'],
                                              labels=True,
                                              sticky=False
                                             )
                       ).add_to(m)

m

Create a styling function with a classification scheme:

import branca

# Create a series (or a dictionary?) out of the variable that you want to map
employed_series = data.set_index('seutukunta')['tyollisyys']

# Setl colorscale
colorscale = branca.colormap.linear.RdYlBu_05.to_step(data = geodata['tyollisyys'], n = 6, method = 'quantiles')

#Define style function
def my_color_function(feature):
    
   employed = employed_series.get(int(feature['id']), None)

   return {
       'fillOpacity': 0.5,
       'weight': 0,
       'fillColor': '#black' if employed is None else colorscale(employed)
       }

Plot the polygons, apply the classification scheme and add tooltips

# Based on :
#https://python-visualization.github.io/folium/quickstart.html
#https://github.com/python-visualization/folium/blob/master/examples/Colormaps.ipynb


# Create a Map instance
m = folium.Map(location=[60.25, 24.8], tiles = 'cartodbpositron', zoom_start=6, control_scale=True)

# add geojson layer on the map, which visualizes the polygons based on your style function, and additional parameters  
folium.GeoJson(
   geodata,
   name = "Employment rate in Finland (2013)",
   style_function=my_color_function,
   tooltip=folium.features.GeoJsonTooltip(fields=['nimi','tyollisyys'],
                                             aliases = ['Region','Employment rate (%)'],
                                             labels=True,
                                             sticky=True)
).add_to(m)


#Add a legend
colorscale .caption = 'Employment rate (%)'
m.add_child(colorscale)

# Create a layer control object and add it to our map instance
folium.LayerControl().add_to(m)

#Show map
m

outfp = "employment_rate_in_finland.html"
m.save(outfp)
Interactive maps Sharing interactive plots on GitHub