If you're a poli-sci student, you've got half the equation.

Programming is useful for a lot of things, but (for me, at least) it really gets exciting when you start to collect and manipulate data. For example, in your field it might be advantageous to collect some data on the current congress's voting habits. From there you realized that you have some interconnected relationships between voting habits and lobbying interests. Well you might as well expand your database and collect some metadata on these corporate interests, who are the share-holders? Who worked where before they were there? Did any of our congress-folk work there as well or have a history with that corporation? Holy-cow! What a bunch of interesting relationships!

You'd use Python to connect to web API's to gather your data, you'd then formulate a model in python to represent the schema of your database and use it to handle your insert, update, and delete logic. Depending on which database you use, there are traditional relational databases MySQL, PostgreSQL, T-SQL and a slew of no-SQL variants like MonogoDB, CouchDB, Cassendra (sounds sexy, right?) would indicate which database API you'd need to read some documentation on to connect to with Python.