Visualization is important and makes our data human readable. It also gives us a quick status check on what is going on with the data and could help us spot issues before they become too large.

We currently have very little visualization - we have a python-based script that can run through CSV files and map them. We also have some GNU Plot based scripts that are still a little clunky to use. We have two approaches that are being considered:

The image based approach means rendering a .png or other image of different sample sets. This approach has the advantage of being very computationally light - rendering work is only done once a day (or however often it needs to be done).

This approach is based around building an API surrounding the database, specifically for data access. This approach might be more interactive and fun, but may be a little more heavy on the database access.

• D3.js (general visualization lib for JS)
  • Codeacademy D3.js tutorial
• cubism.js (library for visualizing time series data)
• Installing Matplotlib
• https://github.com/GraphAlchemist/Alchemy
• https://github.com/plotly/plotly.js

• NY Times baseball strikeouts

Contributing authors:

Created by kluong on 2015/11/04 04:52.