February AWPUG Meetup: Django ORM Adventures and Clustering Log Data
Details
Hey, everyone! This month's meetup will be a treat, since we have two cool talks about different areas of Python web dev. I'm really looking forward to both of these. We also have a cool new sponsor, PromptWorks!
First up, Tom Kane from Atlassian will be delivering "A wild goose chase through the Django ORM." Tom tried to address tech debt in one part of the code by following the advice of a simple 3-year-old comment: "FIXME: we should switch back to using the Django ORM". It was generally accepted that this would happen one day. Little did anyone know that path was fraught with peril. He tried several strategies, caused an incident, revealed a flaw in an open source library, and ultimately replaced the FIXME with a comment explaining why the existing code must remain for performance reasons. Despite not achieving his initial goal, the efforts led to valuable improvements to both Bitbucket and the open source Django Waffle project.
Following that, Jacob Barhak (who runs the Evening of Python Coding meetup) will talk about using ML techniques to cluster web log data. We'll see how Python analytic and visualization tools can be used to analyze web data. For the demo we used an anonymous data set made public by MSNBC through the UCI Machine Learning Repository. The demo code, data, and results are available on https://github.com/Jacob-Barhak/ClusterWebLog. The demo will use Bokeh to visualize the data and scikit-learn to perform basic clustering of the data.
Refreshments will be provided by PromptWorks, who have a new-ish office here in Austin and offered the following: "PromptWorks is a software consulting shop specializing in Ruby, Python, JavaScript and related technologies, with offices in Philadelphia, New York and (of course) Austin. With unmatched technical chops and a serious commitment to Agile practices, we are able to deliver high-quality software with low risk at a predictable pace. In addition to helping businesses build new web apps and APIs, we solve problems with existing apps like scaling, testing and automating development operations. We also build software for machine learning, big data, and DevOps projects."
As usual, we'll be meeting in Capital Factory. If you park at the Omni, remember you can validate your parking with the front desk when you come in. See y'all Thursday!