This project is trying to visualize the most recent twitter activities in NYC hoping to get a picture of trending topics and their geographical distribution. It's a course project for the Information Visualization at NYU Tandon by Professor Bertini.
The project fetches tweets geolocated in NYC, and transmit them to connected browsers. At the client side, a quad tree is maintained to store the hierarchical cluster/word frequency information. This makes tweet grouping automatically adjusted by the zooming area to reach a balance of general/details information.
Screenshot: https://vimeo.com/167764465
Website: http://twitter-nyc.com
Documentation: report.pdf
Prerequisites:
- node.js >=6.0
- the latest npm
- canvas package should be compiled on the host, read the provided document
cd /root/of/project
npm install
npm run build
After that the build
folder will contains all you need to host the server. For deploying on platform such as Heroku, check out the command npm run deploy
, for the server, just copy the folder onto it using command such as scp -r build username@hostname:path
.
npm install
NOTE:
- If you encountered an issue telling you
npm
is killed during the process, make sure your server got at least 1GB RAM. - If you encountered an issue telling you
ndoe-gyp rebuild failed
, please read the canvas package document to install the prerequisites.
Once you finished the installation, just goto build/
folder and run with PORT=80 node server.js
.