Data Diagnostics: missingno

written by Eric J. Ma on 2017-02-06

Sometimes, all that you need is a visual cue on whether the data you have on hand are complete or not. Looking at a table can be dizzying at times, so I'm very glad I found this packaged called missingno! It provides a way to quickly visualize the "nullity" of your dataset. See an example below:

Displaying nullity of a data set.

It's built on top of matplotlib, and takes in pandas DataFrames, which means it plays very nicely with the rest of the PyData stack. I recently took it for a tour when I did a quick stats consult with Mia Lieberman (DCM); the above plot was made using her data, used with permission.

Highly recommended package!