Information Resources Exchange Group
The ACP Building - 8th Floor
190 N. Independence Mall West
Philadelphia, PA 19106
For directions Call (215) 592-1800
Please Use Sixth Street Entrance
1. Call to Order and Introductions
2. Cicero Zip Code to Legislative District Matching
Presented by Aaron Dennis, Data Specialist - Azavea (email@example.com)
Cicero, a product of Azavea, offers solutions for matching constituents with their representatives based on a database of legislative district boundaries. Cicero’s latitude/longitude and address-level matching provide high accuracy matching at the national, state, and local levels, but sometimes users just want to get the job done with a five-digit Zip Code.
Zip Codes are useful because they’re widely adopted, commonly understood, and don’t require personally identifiable information. However, ZIP codes don’t always follow boundary lines like the edges of legislative districts.
We calculated statistics on the accuracy of Zip-to-District matching to see whether we could reduce the error rate with an analytical approach. Using Cicero’s district boundaries database and US Census Bureau population data, we better quantified the issue and developed improved methods. The results show particular significance in the Delaware Valley.
3. My Open Source Software Fellowship at Azavea: Software for Deep Learning on Satellite Imagery
Presented by Yoni Nachmany, Azavea (firstname.lastname@example.org)
2017 marked Azavea’s inaugural Open Source Software Fellowship Program, which matches software engineering fellows to open source projects at Azavea. Azavea is a B Corporation that creates geospatial web software and data analytics for positive civic, social, and environmental impact, while advancing the state-of-the-art through research.
On July 20th, Raster Vision, Azavea’s system for analyzing aerial and satellite imagery using deep learning that works on a variety of computer vision tasks and datasets, placed 23rd out of 938 teams in a Kaggle competition called “Planet: Understanding the Amazon from Space”.
My fellowship focused on three components of software for deep learning: dataset exploration, capacity building, and result demonstration. My talk will highlight my three projects: exploring SpaceNet satellite imagery with a Python geospatial data science stack, building an integration of Raster Vision trained models with the GeoTrellis data processing engine, and demonstrating Raster Vision’s Kaggle results in a web app.
4. GitHub Desktop 101
Presented by Taylor Nation, Operations Engineer - Azavea ( email@example.com)
Learn the basics of how Azavea uses GitHub as a software development tool, and practice your newfound skills by making changes to a sample application. Topics include an overview of Git, making changes to a GitHub repository, using GitHub Desktop, and setting up your own website with GitHub pages. Bring your laptop and Git ready to learn!
5. DVRPC Work Program Update
6. Member Roundtable
Recent Projects - Attendees are encouraged to share recent accomplishments and to discuss any ongoing projects that you or your agency is working on that is related to information gathering and sharing.
Recent Event Reports - Attendees are encouraged to provide a brief summary of any recent meeting or conference they attended.
Upcoming Events - Attendees are encouraged to announce any upcoming meeting or conference.