World Bank Indicators API Example

1. World Bank Indicators API Example#

The following is an example of a Jupyter notebook - a tutorial of how to retrieve data from the World Bank Indicators API - that illustrates how to use computational content with the template.

1.1. Requirements#

import itertools

import pandas
import requests
from bokeh.palettes import Spectral6
from bokeh.plotting import figure, output_notebook, show

1.2. Data Retrieval#

In this example, we retrieve Population, total (SP.POP.TOTL) from the World Bank Indicators for BRICS.

url = "https://api.worldbank.org/v2/country/chn;bra;ind;rus;zaf/indicator/SP.POP.TOTL?format=json&per_page=10000"

Let’s use requests to send a GET request,

r = requests.get(url)

Now, let’s normalize and create pandas.DataFrame from the response,

# normalize
data = pandas.json_normalize(r.json()[-1])

# create dataframe
df = pandas.DataFrame.from_dict(data)

Tip

Alternatively, the World Bank API supports downloading the data as an archive.

Let’s take a look at the dataframe,

df = df.pivot_table(values="value", index="date", columns="countryiso3code")
df = df / 1e6  # scaling
df
countryiso3code BRA CHN IND RUS ZAF
date
1960 73.092515 667.070 445.954579 119.897000 16.520441
1961 75.330008 660.330 456.351876 121.236000 16.989464
1962 77.599218 665.770 467.024193 122.591000 17.503133
1963 79.915555 682.335 477.933619 123.960000 18.042215
1964 82.262794 698.355 489.059309 125.345000 18.603097
... ... ... ... ... ...
2018 210.166592 1402.760 1369.003306 144.477859 57.339635
2019 211.782878 1407.745 1383.112050 144.406261 58.087055
2020 213.196304 1411.100 1396.387127 144.073139 58.801927
2021 214.326223 1412.360 1407.563842 144.130482 59.392255
2022 215.313498 1412.175 1417.173173 144.236933 59.893885

63 rows × 5 columns

1.3. Visualization#

Let’s now plot the data as a time series using Bokeh.

output_notebook()

p = figure(title="Population, total (World Bank)", width=700, height=600)

# colors
colors = itertools.cycle(Spectral6)

# plotting the line graph
for column, color in zip(df.columns, colors):
    p.line(
        df.index,
        df[column],
        legend_label=column,
        color=color,
        line_width=2,
    )

p.legend.location = "right"
p.legend.click_policy = "mute"
p.title.text_font_size = "12pt"

p.xaxis.axis_label = "Year"
p.yaxis.axis_label = "Population, total (in millions)"

show(p)
Loading BokehJS ...
Contents