Why is he getting SO MANY anomalies? | Writing Better Reports with Python & pandas Part 4
Sign up for DUTC Weekly: https://mailchi.mp/dutc/ytweekly
LinkedIn: / do-n-t-use-this-code
Website: https://www.dontusethiscode.com/index...
Upcoming events: learning.dutc.io
Looking for training? Contact [email protected]
This is part of the series Writing Better Reports with Python & Pandas - Getting it right the first time.
Let's discuss core principles behind writing better reports that summarize and communicate your analyses, and we will see how these principles manifest themselves in the tools we use—Python, pandas, and maybe a little Matplotlib.
Our goal is to develop intuitions that will allow us to correctly structure our code to support new user requirements, new tweaks, and requests for (re-)validation and reconciliation.
We'll look to answering the following questions:
how do the basics of Python, like iteration helpers and the itertools module, support better writing of reports?
how do I make sure that I track “provenance”? How does this affect how I model my data and represent inferences?
why is it important for me to track “intentionality” during the data capture process? How does this affect how I write my data pipelines?
how do I make sure that I reduce “entropy” in my report? How does this affect how I make use of tools like pandas to avoid discarding information that might then need to be reconstituted?
why is “bitemporality” a super-power? What can I do when my underlying data sources include all temporal dimensions, and how does this allow to me provide greater confidence in the correctness of my reports?