Next steps for post-hackathon Bonsai


In the Friday morning meeting, Chris gave his ideas for next steps to capitalise on the outputs and momentum of this week. I took some notes (as shared on Slack) and have improved them a littie for sharing here. Comment and suggestions very welcome!

We should reach out to other communities and invite them into the fold: what kind of communities to reach out to, and how best to do it?

With or without Bonsai as official partner
For example: open energy data -> September deadline. H2020 proposal

System Improvement iterations
Assumptions we don't need in Exiobase -> simplify the logic of it
        uniform distribution? No: market. We don't have that information
More general system modelling choices -> average or marginal production values?
Quantifying substitution vs allocation (on specific industries/products?)
Data validation dash-board.
  • here's our model: CO2 emissions, trade etc.
  • here's other data: surveys, satellites, other primary data or model results
  • What's the difference between ours and the others?
Run it daily whilst the data improves -> are we getting closer to "truth"?
New sources of data linked to our model
New inventory data, and scripts for generating them
        e.g. population changes

Document now!
In one week, the fine details will be gone from our minds.
So go back, look at functions, add docstrings
Can we make the ontology more useful or easier?
Which things have we talked about but haven't yet built and uploaded examples for?
To help new people coming in to not get overwhelmed... These "potential useful contributors" need to be able to read and understand the code.
Reproducibility goals aren't yet met by the repository structure yet. Even people following the hackathon aren't all sure. Housekeeping and clear descriptions needed.


Some thoughts on managing projects/working groups:

New project ideas

We need a place where newcomers can browse outstanding or ongoing projects, to get some inspiration or see whether their skills could be put to use. I created a project board for this:, but I am not sure if this is the best layout. It could be better to have a repo called something like projects, where the issues would track each potential or ongoing project, and we could assign and filter by tags. In this repo, each directory would have a more detailed description of the working group/project, including where to find out more information (if applicable).


One of the most important things we can do is to better communicate our current working state, and where others can work on specific tasks. Our BEP on communication is here: One of the things this proposes is to transfer the webopage to a GH repo, and from some research and past experience a static-site generator would work well. The new project board proposes using Jekyll, which makes for trivial hosting on GH pages, and is widely used and supported. It would allow us to use markdown consistently for everything, which is nice (I hear the howls from ReST fanbois - I am one too:). Would someone be interested in dissecting the current site? If you know a bit about HTML and template inheritance, it shouldn't take long (can skip the donation page for now).

One vital thing would be a colorful vertical subway-style map showing how data progresses through the system, and which library is used at each step. Unless you are working with them directly it can be easy to lose track.

BEP voting

The testing strategy for a few BEPs was to test them out at the hackathon. If you are a BEP author or editor, please think about revising your work, or submitting it for a vote.