Learning to learn
To learn about your scientific field you typically look to:
- research papers
- conferences that you attend
- courses provided by your institute
You cannot rely on these sources to learn about effective data analysis!
Many of the most powerful tools for data analysis didn’t - or barely - existed 5 years ago. To know that they exist you need to be regularly watching for new developments. You need to lead your own learning about these developments.
- More or less everything of use can be found in twitter
- But you have to find it amid the rubbish!
- You need to actively search for people who post on the topics you’re interested in
- I’ve included links to people who post useful stuff
- Great for introductory tutorials (e.g. search for bash scripts for beginners)
- Great for conference presentations (e.g. search for scipy 2020)
- Discourse/discord servers
- Increasingly important forums e.g. Dask discourse or Pangeo
- Sometimes forums are private so won’t turn up in search results
- Newletters such as the NotANumber newsletter on high performance python
- Some are very high quality
- Hard to find
- Wide-ranging and in-depth
- Can be expensive
- Go out of date in a few years
- Online courses
- Huge number of courses available
- Hard to find the right one
- Lots of good free courses e.g. FastAI for machine learning
What about Stack Overflow?
- Great resource for ideas
- Provides immediate workarounds
- Generally does not help you understand fundamental issues
What do you think are the most effective ways to learn about data analysis?