Working with European Central Bank Data in Python

Description of the problem

This is my third in my series of working with econometric data queried from a specific data store using a specific API. Today, I am going to work with European Central Bank’s datasets using their API. My language of choice today is going to be python.

Implementation

ECB serves its data in XML format. Python has an excellent library called xmldict that converts an XML object into a python hash map (dictionary in python parlance.) First, let us load the libraries we are going to need.

from urllib.request import urlopen
from xmltodict import parse

Here is the query function. One needs to poke around the hashmap a while to figure out how to extract the relevant bits. It is ugly, to say the least. I am sure ECB has its own reasons for structuring the response this way, but it eludes me.

def QueryECB(series):
    with urlopen("https://sdw-wsrest.ecb.europa.eu/service/data/EXR/" + series) as url:
         raw = parse(url.read().decode('utf8'))
    data = raw['message:GenericData']['message:DataSet']['generic:Series']['generic:Obs']
    res = {x['generic:ObsDimension']['@value']: 
           float(x['generic:ObsValue']['@value'])
           for x in data}
    return res

As an example, I am going to query Euro (EUR) and Canadian Dollar (CAD) daily exchange rate:

res = QueryECB("D.CAD.EUR.SP00.A")

You should poke around this data too. As this is Sunday, and I am still waking up, I am going to just draw it:

import matplotlib.pyplot as plt

dates = res.keys()
xs = list(range(len(dates)))
ys = [res[x] for x in sorted(dates)]

plt.figure(figsize=(12, 6), dpi=80)
plt.ylim(0.0, 2.0)

plt.plot(xs,ys)
plt.show()