NICAR Database Library

Whenever I acquire a new dataset to play with, one of the first things I want to know if how the tables relate to each other. Ideally, I know that for each main record, I'm going to want to know pretty much everything the data says about it. For example, typically when dealing with FEC campaign contributions, I want to know who gave, how much and who they gave it to. And that means joining a few tables, since candidate information is kept separately from donor information.

In the programming world, documents that explain how tables are joined are typically not text-only. Rather, you usually get a schema drawing, showing the tables graphically with lines specifying what fields join together.

Here at the NICAR Database Library, our mission is to make the task of working with databases as easy and straightforward as possible for journalists. So we've begun the task of drawing up schemas for our datasets, starting with a simple one, the IRS Migration dataset.

Schemas will be drawn up in pdf format, and can be downloaded on the page along with the record layout report and record samples.

Our schemas are being drawn by Yan Wang.