Name normalization

[hupili]

Question: Why "郭偉强" appeared twice in the plot?

Reported for commit: 76b7826cabac4ce9a9f6233f57c92184cdd76d4c Fixed in commit: 8b95798aa51ac512c4eb8d0ed98cbee12e5257c4

Kenneth Chen Wei On: Case of Hon KWOK Wai-Keung's Chinese name is a good case example of data maintenance. When he changed his name from 郭偉強 to 郭偉强 last year, we should have considered the implication of the said change to our data records. Good catch!

Pili Hu: Kenneth Chen Wei On , also 'Dr Joseph LEE' and 'Prof Joseph LEE'. I had this cleansing step in the notebook, http://bit.ly/1k58I8b I think it would be better to officially give the members an ID. Then we do not have data normalization issue. The same for motion ID. Motion is easier because we can use date+order as ID. Name can have some subtle issues.

Refs:

• https://www.facebook.com/kchenweion/posts/280751805419946

[hupili]

This is a demo. The ingredients I'm looking for:

• Self-contained. Although the issues are not together with the repo, there exists tools to backup every aspect of a GitHub repo. This makes our work safe, especially if I'm going to type 500 words to explain something.
• Associate commit numbers where the problem is observed and where the problem is fixed. This is part of reproducible research and distinguishes us from many other sources.
• Have reference pointers so that people can trace to Facebook, Disqus, etc. The data hackers can focus on their core business: ensure the technical correctness. Community curators can selectively sync important information to/from those places.
• No matter the issue is open or closed, people can find it. I think the refinement process is a good learning material for everyone. We don't want to say: here's the ultimate unbeatable report, take it. We want to show the methodology, the problems and solutions.