Often developers get too attached to the code that they write.  So much so that we dread losing it.  But, what happens when you intentionally delete code and rewrite it?  You might be surprised at the result.

Using flask, websockets, and redis to make a simple multi-user drawing surface in D3.

I am Director of Engineering for Discovery Communications Emerging Business and Strategy team. We just relaunched Curiosity.com. We use Python all over the place along with MongoDB/Redis/ElasticSearch The site lives within AWS utilizing several of their services. Including EC2, ELB, Route53, Cloudwatch, S3 I would like to discuss our overall architecture and our use/love of Python. And answer any questions on architecture/scalability/process/code.

Brian will talk about his experiences using Python and NLTK http://nltk.org/ to run language comparisons to generate lexical difference graphs like the one mentioned in the "Lexical Distance Among the Languages of Europe" article. http://bit.ly/1cS46Ba The focus will be on the NLTK and how its internals work to process a language. This talk will be his best one ever.

Reference counting is very useful but it has an odd problem. We employ a technique from graphs to approach it. The solution works but it's a bit slow.

Seen on garish LED roadway signs all around Chicago on New Year's Eve, 2013: 986 TRAFFIC DEATHS IN 2013. It leads to many questions: On what roads? When did the accidents happen? What do we do now? I'm scared to drive. I will talk about purging my fears by finding the data to answer some of those questions. http://tothebeat.github.io/fatal-car-crashes/ This talk will involve PythonAnywhere, IPython, a module that's not even on PyPi (dbfpy), searching for and finding open government data, CartoDB, Google Fusion Tables, csv, and maybe Pandas. Rest assured, there will be no graphic photos.

Braintree needs to be highly available and secure, while still maintaining a rapid development pace and strict backwards compatibility. In order to achieve that, we use what has become known as the "Chicago Process". This involves pairing, strict TDD, a team structure, and weekly iterations, all to empower the devs to make decisions and get work of a high quality done while avoiding siloing.

Pandas is the data-munging Swiss Army knife of the Python world. Often you know how your data should look but it's not so obvious how to get there, so I'll present a visual approach to learning the library and data manipulation.