cooccurrences

Counts of pairs of places appearing in same Wikipedia article

Tab-Separated Values Files

You can download the counts as a zipped tsv from here. The tsv looks like this:

a	b	count
New York City	United States of America	500
Russia	Soviet Union	300

Pickled Counter

You can also download the counts as a zipped pickled Python Counter from here. Here's how you can use it:

import pickle

with open("/tmp/cooccurrences.pickle", "rb") as f:
    cooccurrences = pickle.load(f)

count = cooccurrences[("New York City", "United States of America")]
print("New York City and United States of America appear together " + str(count) + " times.")

Sorted

The keys are sorted using Python's sorted method. You must sort your keys before looking up the results. For example:

two_random_keys.sort()
a, b = two_random_keys
count = cooccurrences[tuple(two_random_keys)]
print(a + " and " + b + " appear together " + str(count) + " times.")

Useful for Machine Learning

You can use this data to decide what cooccurrence of places makes the most sense. For example:

a	b	count
Azerbaijan	Georgia (country)	2033
Azerbaijan	Georgia (U.S. state)	101

Contact

[email protected]