What will we cover in this tutorial?
In this tutorial we will visualize the inflation on a map. This will be done by getting the inflation data directly from World Bank using the Pandas-datareader. This data will be joined with data from GeoPandas, which provides a world map we can use to create a Choropleth map.
Step 1: Retrieve the inflation data from World Bank
The Pandas-datareader has an interface to get data from World Bank. To find interesting data from World Bank you should explore data.worldbank.org, which contains various interesting indicators.
When you find one, like the Inflation, consumer prices (annual %), we will use, you can see that you can download it in CSV, XML, or excel. But we are not old fashioned, hence, we will use the direct API to get fresh data every time we run our program.
To use the API, we need the indicator, which you will find in the url. In this case.
https://data.worldbank.org/indicator/FP.CPI.TOTL.ZG
Hence we have it FP.CPI.TOTL.ZG.
Using the Pandas-datareader API you can get the data by running the following piece of code.
from pandas_datareader import wb
data = wb.download(indicator='FP.CPI.TOTL.ZG', country='all', start=2019, end=2019)
print(data)
If you inspect the output, you will see it is structured a bit inconvenient.
FP.CPI.TOTL.ZG
country year
Arab World 2019 1.336016
Caribbean small states 2019 NaN
Central Europe and the Baltics 2019 2.664561
Early-demographic dividend 2019 3.030587
East Asia & Pacific 2019 1.773102
East Asia & Pacific (excluding high income) 2019 2.779172
East Asia & Pacific (IDA & IBRD countries) 2019 2.779172
It has two indexes.
We want to reset index 1 (the year) and, which will make year to a column. Then for convenience we should rename the columns.
from pandas_datareader import wb
data = wb.download(indicator='FP.CPI.TOTL.ZG', country='all', start=2019, end=2019)
data = data.reset_index(1)
data.columns = ['year', 'inflation']
print(data)
Resulting in the following.
year inflation
country
Arab World 2019 1.336016
Caribbean small states 2019 NaN
Central Europe and the Baltics 2019 2.664561
Early-demographic dividend 2019 3.030587
East Asia & Pacific 2019 1.773102
East Asia & Pacific (excluding high income) 2019 2.779172
East Asia & Pacific (IDA & IBRD countries) 2019 2.779172
Step 2: Retrieve the world map data
The world map data is available from GeoPandas. At first glance everything is easy.
import geopandas
map = geopandas.read_file(geopandas.datasets.get_path('naturalearth_lowres'))
map = map[map['name'] != 'Antarctica']
print(map)
Where I excluded Antarctica for visual purposes. Inspecting some of the output.
pop_est continent name iso_a3 gdp_md_est geometry
0 920938 Oceania Fiji FJI 8374.00 MULTIPOLYGON (((180.00000 -16.06713, 180.00000...
1 53950935 Africa Tanzania TZA 150600.00 POLYGON ((33.90371 -0.95000, 34.07262 -1.05982...
2 603253 Africa W. Sahara ESH 906.50 POLYGON ((-8.66559 27.65643, -8.66512 27.58948...
3 35623680 North America Canada CAN 1674000.00 MULTIPOLYGON (((-122.84000 49.00000, -122.9742...
4 326625791 North America United States of America USA 18560000.00 MULTIPOLYGON (((-122.84000 49.00000, -120.0000...
5 18556698 Asia Kazakhstan KAZ 460700.00 POLYGON ((87.35997 49.21498, 86.59878 48.54918...
6 29748859 Asia Uzbekistan UZB 202300.00 POLYGON ((55.96819 41.30864, 55.92892 44.99586...
It seems to be a good match to join the data on the name column.
To make it easy, we can make the name column index.
import geopandas
map = geopandas.read_file(geopandas.datasets.get_path('naturalearth_lowres'))
map = map[map['name'] != 'Antarctica']
map = map.set_index('name')
Step 3: Joining the datasets
This is the fun part of Data Science. Why? I am glad you asked. Well, it was an irony. The challenge will be apparent in a moment. There are various ways to deal with it, but in this tutorial we will use a simplistic approach.
Let us do the join.
from pandas_datareader import wb
import geopandas
pd.set_option('display.width', 3000)
pd.set_option('display.max_columns', 300)
pd.set_option('display.max_rows', 500)
map = geopandas.read_file(geopandas.datasets.get_path('naturalearth_lowres'))
map = map[map['name'] != 'Antarctica']
map = map.set_index('name')
data = wb.download(indicator='FP.CPI.TOTL.ZG', country='all', start=2019, end=2019)
data = data.reset_index(1)
data.columns = ['year', 'inflation']
map = map.join(data, how='outer')
print(map)
Where I use an outer join, to get all the “challenges” visible.
Russia 1.422575e+08 Europe RUS 3745000.00 MULTIPOLYGON (((178.72530 71.09880, 180.00000 ... NaN NaN
Russian Federation NaN NaN NaN NaN None 2019 4.470367
...
United States NaN NaN NaN NaN None 2019 1.812210
United States of America 3.266258e+08 North America USA 18560000.00 MULTIPOLYGON (((-122.84000 49.00000, -120.0000... NaN NaN
Where I only took two snippets. The key thing is here, that the data from GeoPandas, containing the map, and data from World Bank, containing the inflation rates we want to color the map with, are not joined.
Hence, we need to join United States together with United States of America. And Russia with Russian Federation.
We would use a location service, which maps counties to country codes. Hence, mapping each data sets country names to country codes (note that GeoPandas already has 3 letter country codes, but some are missing, like Norway and more). This approach still can have some missing pieces, as some country names are not known by the mapping.
Another approach is to look find all the data not mapped and rename them in one of the datasets. This can take some time, but I did most of them in the following.
from pandas_datareader import wb
import geopandas
map = geopandas.read_file(geopandas.datasets.get_path('naturalearth_lowres'))
map = map[map['name'] != 'Antarctica']
map = map.set_index('name')
index_change = {
'United States of America': 'United States',
'Yemen': 'Yemen, Rep.',
'Venezuela': 'Venezuela, RB',
'Syria': 'Syrian Arab Republic',
'Solomon Is.': 'Solomon Islands',
'Russia': 'Russian Federation',
'Iran': 'Iran, Islamic Rep.',
'Gambia': 'Gambia, The',
'Kyrgyzstan': 'Kyrgyz Republic',
'Mauritania': 'Mauritius',
'Egypt': 'Egypt, Arab Rep.'
}
map = map.rename(index=index_change)
data = wb.download(indicator='FP.CPI.TOTL.ZG', country='all', start=2019, end=2019)
data = data.reset_index(1)
data.columns = ['year', 'inflation']
map = map.join(data, how='outer')
Step 4: Making a Choropleth map based on our dataset
The simple plot of the data will not be very insightful. But let’s try that first.
map.plot('inflation')
plt.title("Inflation 2019")
plt.show()
Resulting in the following.
A good way to get inspiration is to check out the documentation with examples.
Where you see a cool color map with scheme=’quantiles’. Let’s try that.
map.plot('inflation', cmap='OrRd', scheme='quantiles')
plt.title("Inflation 2019")
plt.show()
Resulting in the following.
Adding grey tone to countries not mapped, adding a legend, setting the size. Then we are done. The full source code is here.
from pandas_datareader import wb
import geopandas
import pandas as pd
import matplotlib.pyplot as plt
map = geopandas.read_file(geopandas.datasets.get_path('naturalearth_lowres'))
map = map[map['name'] != 'Antarctica']
map = map.set_index('name')
index_change = {
'United States of America': 'United States',
'Yemen': 'Yemen, Rep.',
'Venezuela': 'Venezuela, RB',
'Syria': 'Syrian Arab Republic',
'Solomon Is.': 'Solomon Islands',
'Russia': 'Russian Federation',
'Iran': 'Iran, Islamic Rep.',
'Gambia': 'Gambia, The',
'Kyrgyzstan': 'Kyrgyz Republic',
'Mauritania': 'Mauritius',
'Egypt': 'Egypt, Arab Rep.'
}
map = map.rename(index=index_change)
data = wb.download(indicator='FP.CPI.TOTL.ZG', country='all', start=2019, end=2019)
data = data.reset_index(1)
data.columns = ['year', 'inflation']
map = map.join(data, how='outer')
map.plot('inflation', cmap='OrRd', scheme='quantiles', missing_kwds={"color": "lightgrey"}, legend=True, figsize=(14,5))
plt.title("Inflation 2019")
plt.show()
Resulting in the following output.
Can we get inflation rates by month?
Thanks!