This is an introduction to / cookbook for Pandas, Python's best known data analysis toolkit, built on the foundation laid by Numpy and MatPlotLib.
When trying to learn a new library, I usually find the examples / tutorials to focus slightly too much on the end result and not enough on the process and various tricky parts on which it is easy to get stuck.
This series is therefore more verbose and hopefully will help people (like me at least) to grasp powerful and feature-rich library such as Pandas.
I've used Pandas only once to help a friend in need and only quickly so this was also my real introduction to Pandas, which highlights its ease of use. I'm kinda experienced doing this kind of data wrangling and manipulation, so your mileage may vary
This is not an exhaustive "getting started guide" but rather a focused (CSV file, Time Series, Plotting, Grouping, Sampling, Selecting) walkthrough using data from Dublin's Emergency / Fire Brigade, specifically this file, freely available for download.
-
All of it was done through Ipython / Jupyter interactive notebook software and if you want to run those notebooks, you'll need to install it. It's a great tool for this kind of work and I personally use them a lot.
-
If you're on Windows, some steps might not work (%timeit comes to mind) but they're not mandatory. It's also best to use Anaconda since installing some Python libraries on Windows can be tricky
-
Reading in the appropriate order is strongly recommended. And so is perusing the official documentation and cookbooks.
If you spot any mistake or simply need to get in touch, I'm on Twitter (don't use it that much but I'm still getting notifications :) @lapnitlenav or LinkedIn
Hope this helps you harness the Pandas.
Best, Val.