To let users try out juttle use cases easily, the proposal is to have a setup like this:

• examples/README.md about the demos
• examples//README.md explaining the purpose of the demo, and steps to follow
• examples//dockerfile to stand up the required backend
• examples//data.json with data to be loaded into the backend (unclear: would the docker container do the loading? or as an explicit step by the user?)
• examples//*.juttle - juttle programs that work with the dataset

The dockerizing work is a separate issue #20.

Need to validate that this is the right approach. Also, figure out the testing aspect; will the demo itself serve as an end-to-end test? What automation makes sense?

For the set of demos, let's start with what we already prepped:

• elastic with memes data set - move from private - see PR Adding demo of elastic adapter with 2009 memes dataset #22
• postgres with disk stats - move from private
• influx with cadvisor
• twitter race
• stock ticker via http from yahoo finance
• move the 3 above from private
• github tutorial - see PR First draft of the juttle tutorial with github data juttle#15 with tutorial text, and PR Demo examples: adding dataset and programs for the juttle tutorial #21 with data+juttles

Once we get one example up, will make separate issues for each of the others.