Localize trends in NYC traffic injuries

Our guest speaker and domain expert will be Hassan Adekoya, IT/GIS Director at the NYC Department of Transportation.

We are introducing a new way to deploy computational apps, with an emphasis on spatial and temporal input and output. In this hands-on workshop, participants will work in groups to configure and deploy a geospatial web application to visualize trends in NYC traffic injuries, courtesy of datasets provided by the NYC Department of Transportation. We will cover the basic algorithms behind the application, including how to find nearest neighbors using kd-trees, select an appropriate optimization function and forecast values of a time series. We will also spend the first ten minutes reviewing highlights from the NIPS 2015 conference (http://nips.cc/). Please bring your laptop.

Amazon Web Services (https://twitter.com/AWSstartups) is graciously providing a venue for our event. Thank you to Leilani at the NYC AWS Pop-Up Loft and to Jason at Amazon for making this possible.

Community health and safety topics
• Visualize seasonal trends in NYC traffic injuries

Algorithms
• Find nearest neighbors using kd-trees
• Select optimization techniques
• Forecast values of a time series

Operations
• Write a configuration file to configure a script as an app
• Run the app locally
• Serve the app locally

Agenda
6:30pm Review highlights from NIPS 2015 Montréal (http://nips.cc/)
6:40pm Learn from Hassan Adekoya, IT/GIS Director, NYC Department of Transportation
6:50pm Split into groups and meet each other
7:00pm Guide participants through steps and concepts
7:10pm Help participants as they work through the tutorial

Code and walkthrough
https://github.com/crosscompute
https://pypi.python.org/pypi/crosscompute/

Datasets
http://www.nyc.gov/html/dot/html/about/vz_datafeeds.shtml

Tips on preparing for winter storms
http://m.fema.gov/winter-storms-extreme-cold