Past Meetings • Recent Topics

Expanding Our Early Intervention System for Adverse Police Interactions
(0:20:00 Minutes)
By: Sumedh Joshi, Jonathan Keane, Joshua Mausolf, Lin Taylor

Many police departments in the United States use “early intervention systems” to identify officers who may benefit from additional training, resources, or counseling. These systems attempt to determine behavioral patterns that predict a higher risk of future adverse incidents, ranging from excessive use of force and citizen complaints to on-duty accidents and personal injury. Detecting these risk factors enables departments to develop targeted interventions and make operational changes to protect officer safety and improve police/community interactions. Last summer, DSSG worked with the Charlotte-Mecklenburg Police Department on building a better early intervention system, applying data analysis to provide insights on individual and situational risk factors for adverse interactions. This year, we will partner with additional police departments, including the Metro Nashville Police Department, to test and expand this work in new municipalities, improving both the overall model and local performance. Like last year, we will use anonymized police data and contextual data about local crime and demographics to detect the factors most indicative of future issues, so that departments can provide additional support to their officers.

Predictive Enforcement of Pollution and Hazardous Waste Violations in New York State
(0:20:00 Minutes)
By: Jimmy Jin, Maria Kamenetsky, Dean Magee

New York State’s Department of Environmental Conservation (NYSDEC) is the regulatory agency for environmental issues in the state. Their mission is to conserve, improve and protect New York State’s natural resources and environment and to prevent, abate and control water, land and air pollution. NYSDEC currently conducts approximately 700 inspections each year of facilities in the state that manage hazardous waste. DSSG will work on more effectively allocating inspection resources by creating predictive models that identify facilities with high likelihood of violating environmental regulations. In 2015, we worked with the federal EPA targeting hazardous waste facilities subject to the Resource Conservation and Recovery Act (RCRA). With inspection data from NYSDEC and the public RCRA dataset, we will build a similar model to identify RCRA violators specifically in the New York region, as well as further explore the possibility of applying models to other compliance inspection programs, such as the Clean Air and Clean Water Acts.

Intro to Deploying Django with Ansible
(0:15:00 Minutes)
By: Joe Jasinski

Deploying Django is a breeze when using Ansible. Learn a bit about the power that Ansible provides and how easy it is to get started using it!

Getting meaningful results from unit tests
(0:10:00 Minutes)
By: Manu Phatak
Slides Link
We count on unittests failure messages to give us reasonable feedback on how to proceed with a failing test. When you're working with builtin data structures and objects, unit test feedback is usually pretty good. It helps you identify and solve the problem. In contrast, the default results you get from custom objects can be practically useless. https://gist.github.com/bionikspoon/2e434a2c193a06b0996cc98c6a604de9

Computer Vision in Python: How to build a basic face detection system
(0:25:00 Minutes)
By: Jeremy Watt

In this short talk Jeremy will describe the universal pipeline for performing object detection (that is, the automatic detection of objects in digital images) in Python. This will include a discussion of various classification schemes, feature extraction methods, and their fusion in the form of deep neural networks. Demo code illustrating these concepts will be shown using the IPython notebook environment.

JIRA + Python
(0:15:00 Minutes)
By: Jonathan Pietkiewicz

JIRA is a popular issue tracking and project management software. In this talk you will learn about JIRA and how to interact with the tool using the jira-python library. Primary topics covered will include an overview of the API, creating and modifying issues, linking issues, and searching from Python.

PyCon 2016 recap
(0:15:00 Minutes)
By: Jerry Dumblauskas

Let's give an overview of Pycon

An overview of python projects of OS X administration
(0:20:00 Minutes)
By: Ryan Manly

In this talk Ryan will give a brief overview of several python projects used by many OS X admins to provide cached update services, imaging, software deployment, and configuration management. Looking at the code in these tools can provide insight into Apple's preferences frameworks and if you are in a "DevOps" type role some of the projects discussed may help you immensely.

Python Hype?
(0:30:00 Minutes)
By: Brian Ray

Brian will reveal the survey results that will help shed some light on the current and future projectile of the Python programming language. Has Python reached a peak? Will it popularity continues to rise? What are the users of Python at different levels saying about the state of Python Programming Language?

Python, Startups, Tech Debt, and You
(0:20:00 Minutes)
By: Matt Erickson

There's a lot of people newish to Python and either interested or already in a startup environment (either within a larger corporation or an actual startup). Python makes a *great* tool for that, however while there’s ways to use it to carry your work along to great success, there’s ways to wind up with such spaghetti you’re tempted to throw your hands in the air and go back to Java. The focus on the talk is how to use Python and the tools it provides to avoid the unmaintainable mess while still being able to “cut corners” to get something out the door to make your boss/investors/customers happy.

pyStan: Bayesian Inference for Fun and Profit
(0:30:00 Minutes)
By: Stephen Hoover
Slides Link
Probabilistic programming languages offer a flexible and expressive way to model data by treating random variables as first-class objects. Stan is a popular and well-supported library which allows users to write models in the Stan programming language and use MCMC methods to perform Bayesian inference. Stan itself is written in C++, and has a Python interface through the PyStan package. In this talk, I'll show off some of the capabilities of PyStan and go through a simple practical example of Bayesian inference in Python.

The 'collections' module
(0:10:00 Minutes)
By: Phil Robare

A quick overview of the collections module and its five data structures. The talk will be aimed at the intermediate level python user who has the basic syntax down but has not yet delved into the wealth of programming tools in the standard library.

Multiple System Failure: A case study in debugging
(0:30:00 Minutes)
By: Adam Forsyth

Recently, the Braintree Python library wasn't working on Google App Engine. Braintree, GAE, requests, and urllib3 all had problems and I tracked down each one. I'll walk you through debugging with only basic tools -- editing the code to observe state and using git to find the responsible commit. This talk expects a basic understanding of web programming, git, and Python.

Hacking Bokeh
(0:20:00 Minutes)
By: Brian Ray

I brief introduction into Bokeh http://bokeh.pydata.org/en/latest/ And a bit on how to build interactive graphs in jupyter notebooks or stand alone.

module of the month - usaddress
(0:10:00 Minutes)
By: Cathy Deng

usaddress is a python library that uses NLP methods to parse address strings into structured components (e.g. street name, city, zip). it is trained on real-world addresses with real-world data quirks - as a result, it's robust in handling messy data. usaddress was built by DataMade, a local civic technology company. TL;DR usaddress helps you avoid regex for address data, which is a terrible rabbit hole.