Applied Data Science — Handwritten Character Recognition — Part I of IV

(no, this will NOT involve the classic MNIST data set . . . .)

Our data for November and December is a simple entry point to image analysis: automated recognition of historical, handwritten, Chinese characters, some of which may date back to over 3,000 years ago.

No prior knowledge of Chinese languages is necessary.

Some of the questions participants may choose to investigate include:

• Is there a way to reliably identify what timeframe a character was written in?
• Are there relationships between words within different Chinese scripts?
• Are there patterns of similarity between images of the same word?
• Is it possible to automatically convert ancient handwriting to modern characters?

Everyone from beginners to experts is welcome. As always, our goals are:

1. Provide skill-level-appropriate challenges to all participants.
2. Participants work with and learn from one another.
3. Those who attend all four Sundays finish a project they can add to their github/resume/portfolio.

Bring your favorite data analysis software! We're language-agnostic

If you don't already have a favorite analysis tool, here are a few suggestions: http://goo.gl/W8Q6cV