Few of the students grouped at the podium in the front of the General Assembly classroom had any prior experience as first responders to a hurricane, but, despite that, they clicked through slides outlining innovative and data-driven ways to respond in the event of one. Through a collaboration between General Assembly and New Light Technologies (NLT), each of the three groups was tasked with developing a tool aimed at saving the most lives in the event of a catastrophic storm using cutting edge data science tools and spatial analysis. The presentation was the capstone of their twelve-week immersive program in data science and visualization, with the addition of the geographic lens that NLT brings to all its work.
General Assembly is a school that focuses on developing its students’ technical and coding skills — offering everything from one-day workshops on specific skillsets to full-time programs in data science, web development, and user-experience design. Its classes are structured for people from all walks of life who may be looking to pivot into an entirely different career or maybe looking to deepen their expertise in an existing one.
Today, General Assembly’s immersive Data Science class is poised to showcase their two-week capstone project to a room of instructors and industry leaders. It’s the final leg of their journey through a incredibly rigorous and concentrated skills training programming requiring 480 hours of in-seat instructional and lab time. (To compare, a master’s program comprises about 460 hours!)
NLT, a contractor that works closely with the Federal Emergency Management Administration (FEMA) providing GIS and IT services, analytics and remote sensing analysis, provided the students with real questions facing its GIS work in the field of emergency preparedness and response. “We challenged the students to find innovative solutions that would help disaster support agencies know in a timely manner the location of people who need help, the location of communities that are not connected to electricity or cannot access medical care,” Ran Goldblatt, Remote Sensing Scientist & Senior Consultant at NLT, said of the partnership with General Assembly.
The groups presented on three distinct topics: including leveraging social media to map disasters; using live feeds to determine road closures during an emergency; and extracting building values from Zillow to determine property damage estimates during a storm. The group members would use the tools and skills they’ve been developing throughout their intensive twelve-week education. One group ran nearly 10,000 tweets through an algorithm that performed a sentiment analysis on each one, and then spatially visualized the presence of urgent and non-urgent tweets to show where people might need the most help after a disaster. The next group would parse data using Python functions to define important words; to extract text features from data; to score words using an algorithm to parse relevant text, and to break text into words from tweets, news headlines and FEMA articles – the end product being a database of keywords with which you can search for event hashtags, geographic location, and road intersections. Another developed a dynamic code that calls Zillow’s application programming interface (API) directly and pairs it with additional coefficients to estimate the monetary cost of certain catastrophic events.
“We’ve done a number of projects over the last three months,” one of the presenting students Jamila Holt described. “We started off designing a video game. We worked with predicting housing prices in different states around the country. We’ve looked at predicting diseases based on certain factors. We’ve done a time-series analysis using financial data. We’re always doing something a little bit different each week, so to shift to working with FEMA and New Light with mapping was not a hard shift.”
The room was full of GA instructors, NLT researchers, and industry professionals. When each group finished discussing the content of their work, the question-and-answer segments burst with audience interest, and afterward, invitations were extended for students to discuss their work again with at the FEMA office itself.
“This is only the beginning of our partnership with General Assembly,” Goldblatt said as he looked toward the future of further collaboration. “I hope we will help General Assembly incorporate geospatial big data in their data science curriculum and eventually play an important role in training the next generation of data scientists to come.”