Things to do at the hackathon

[tracykteal]

For the hackathon there are few infrastructure and development things to do, as well as going through the lessons to check on and update formatting and make sure things work and make sense. Below are outlined activities for the hackathon.

Infrastructure and development

1. work on introduction to the workshop content Move much of this content to 'data exploration' lesson introduction-RR-Jupyter#34
2. look into slide strategy, write up a how to, so we can use it for all slide decks Slide deck format workshop-RR-Jupyter#2
3. lesson on using git in Publication lesson add lesson on github publication-RR-Jupyter#34
4. lesson on mybinder in Publication lesson Add a section on mybinder publication-RR-Jupyter#33
5. template Jupyter notebook for episodes templates for this workshop workshop-RR-Jupyter#3
6. update the workshop landing page with links to rendered lessons Update page with any lesson content updates workshop-RR-Jupyter#7
7. Workshop setup instructions and plan for running Jupyter without a local installation

Updating individual lessons

7. Data and project organization
8. Data exploration
9. Automation
10. Publishing

For each lesson, we should do the following:

• update index page of the lesson
• use Jupyter notebook lesson template for Jupyter notebook episodes
• check and update format of Markdown episodes
• go through issues that have been filed for each lesson
• implement slide strategy where needed
• run through the lessons and see if they make sense. Make updates as needed. Particular attention to exercises.

[tracykteal]

All of these things can be done in parallel, although we should get started on all the infrastructure and development items at the beginning of day 1. Some things also won't take the whole time, so people can work on more than one thing.

In this issue, please add other things you think we need to do and identify the project or lesson (1-10) or one of the new ones that you'd like to start with!

1-2 people can be on each project. If you have a chance to review the relevant issue or content before the hackathon that would be great!

Some people may be doing hackathon things remotely, so also indicate if you're going to be local or remote, so we're sure to loop everyone in.

[hlapp]

@kcranston @burkesquires @dleehr @JamiesHQ any things worth adding from the post-mortem after the March 2017 workshop at Duke?

One thing I remember is that the feeling was (too) much precious time was spent on the first day outside of Jupyter Notebook, and that it would be good looking for ways to be in Jupyter Notebook early on and more often. I think this concerned especially the intro and the data & project organization parts.

[kellieotto]

Happy to work on #6 in infrastructure and work on the Data Exploration lesson. It's the one I worked on at last year's hackathon and it still has quite a few open issues.

[tracykteal]

Other things to do are to plan for teaching and maintenance:

• Plan more workshops, including see who would be interested in teaching where
• Maintainers for lessons (we'll recruit more, but would be good to see who at the hackathon is interested in what lessons)
• Curriculum committee
• Planning a release date for the curriculum

[choldgraf]

I'm planning to work on the mybinder lesson, as well as to chat about any other ways we could integrate Binder into the materials for the course!

[choldgraf]

Also - a quick question for everybody (slightly tangential), would people be willing to spend about 20-30 minutes as a quick breakout session to brainstorm ideas for how we can improve the tooling around auto-grading in the data science world? We have a few tools in the grading space but many of them haven't quite hit the sweet spot of modularity / functionality. I'd be interested in hearing from you all what kinds of things would make up the most useful open-source grading tool.

[snacktavish]

I will be late (arriving sometime Tuesday afternoon) but I plan to focus on the automation section.

[naupaka]

@choldgraf I’d definitely be interesting in a discussion on auto-grading

[choldgraf]

Two points worth discussing:

• To what extent do we want to really focus on reproducibility + Jupyter-specific things, as opposed to more general topics like data exploration etc.
• How do we want to define "Jupyter"? Just notebooks? There are a lot of other pieces of the Jupyter world (e.g. jupyter lab, jupyter hub) that make inroads with reproducibility etc.