Wondering where to find reliable datasets you can use to do #ddj? Quartz’s data editor, Chris Groskopf, just released a curated list of useful datasets published by important sources.
The point of this directory is to help you find data, which doesn’t necessarily mean the data is easy or simple to use. Also, it is important to have in mind that "reliable" does not mean "correct", thus you will need to understand the data and treat it like you would do with any human source.
The directory includes close to 80 different sites where you can find useful data. Even though the list focuses on the United States of America, it includes at least 30 sites that publish global data.
The data included in the directory was chosen because it was produced by an organization with the statistical capacity and domain knowledge; that it has generally agreed to be unbiased (at least within the context of the data); that is actively updating data on regular intervals; must be likely to be used in the course of routine reporting, and must contain a well-defined sample.
You can access the dataset here and share it freely.