This weekend’s PyGotham conference renewed an appreciation for all things Python in NYC. Python is a programming language known for being concise and easy for humans to read. [As the R-using sole holdout at Harmony Institute, I’m still getting used to Python’s conventions and approaches to data gathering, analysis, and presentation. PyGotham was a chance for me to see what others are doing, and get excited about new possibilities.]

The range of topics, industries, and goals represented by different users at PyGotham was impressive. From detecting sarcasm in speech to developing a database of NYC boiler information, presenters discussed how they’re using the seemingly infinite flexibility Python to attack their problems. The audience also had a diverse range of expertise required, from the newer users wrapping heads around constructing a web app (Thanks, Kat Chuang) or a data pipeline (Thanks, Joe Cabrera), to more experts debating the relative merits of running applications on cloud container services (Thanks, Jeff Uthaichai and Chris Becker) and scaling automated recommendations in real time (Thanks, Brian Muller).

A constant theme throughout the weekend was the uphill battle of data organizing. Data is everywhere, but generating real insight from any dataset demands a lot of specific effort. At PyGotham, much discussion swirled around working with city and federal governments’ woefully inaccessible data and out of touch systems. Several presenters reflected on the chaos of receiving different, idiosyncratic file formats; approaches to extracting data from PDFs came up repeatedly. While it’s not unique to Python users, the Python packages and solutions available for scraping, organizing, and using data are sophisticated. Finding ways around these stumbling blocks was also cause for celebration — presenters spoke about the package BeautifulSoup with reverence, and applause broke out for Steven Slotterback’s workaround for grabbing NYC boiler data without having to fill out a captcha every time.

 

---

 

To learn more about Python and the specific topics discussed,you can find recorded conference talks at http://www.pyvideo.org/. Special thanks to the Women in Machine Learning and Data Science group, who extended the invitation!