The challenges posed by JP Morgan and ArcGIS
What it does
CCCA, or Crowdsourced Cartographic Catastrophe Analysis, is a powerful tool for social good that leverages real-time data from social media. It uses the frequency of tweets with location enabled and keywords relevant to natural disasters to form a heatmap from which first responders and helpers can decide which areas are most in need of assistance.
How we built it
The underlying algorithm, written in Python, uses the Tweepy library to extract location data, if available, from tweets that contain one or more keywords pertaining to natural disasters. A heatmap is created by plotting these points on a minimalistic representation of the globe to keep the focus on the data. The sample size becomes large enough as more information is collected that it is easy for the user to determine real areas of concern by the patterns evident on the map.
Challenges we ran into
Timeout issues with Tweepy, formatting data correctly, automating conversion from CSV to Shapefiles, navigating a new API.
Accomplishments that we're proud of
Streaming a custom, constantly updating data set from Twitter rather than using a preexisting data set, as well as creating a highly readable minimalist UI.
What we learned
How to stream data from Twitter using Python, how to represent spatial data in an impactful way using ArcGIS.
What's next for CCCA (Crowdsourced Cartographic Catastrophe Analysis)
Using machine learning to filter out noise and make sure that the maps only represent meaningful data, finding nearby emergency response services and alert them when help is needed, making a Twitter bot that would tweet out the location and severity of natural disasters, modifying the Twitter streaming algorithm to create maps of other types of events.