Motivation (aka what most of us already know)
Future energy networks will become increasingly tightly coupled: gas could be generated from renewable electricity and fed into the gas network for heating, power production, and transport fuels. Under-utilised existing gas network assets may provide storage capacity for renewable energy, serve as alternative means of transporting renewable energy, and potentially facilitate energy imports from outside Europe. So it might be of value including gas network operation/utilisation and investments in energy system models.
I think it is worth following up on this breakout group from 2 years ago.
This part of the do-a-thon will investigate retrieving grid data on transmission routes and capacities from official sources (e.g. ENTSO-G). It could e.g. look at whether the GridKit tool  – previously used to generate an extract of the ENTSO-E interactive map  – can be generalised to also make an extract of the ENTSO-G transparency map ; starting off with identifying relevant components and their attributes required to build a gas grid model. Personally, I would also be interested to learn about the status of the SciGRID_gas  project initiated by the DLR Institute for Networked Energy Systems, in case somebody who is involved participates.
If time allows, this part of the do-a-thon will investigate modelling the flows in a gas transmission network. Existing open-source modelling tools (e.g. GasModels.jl  by LANL-ANSI) have nonlinear mixed-integer gas flow formulations. To retain the continuous and linear/convex nature of most energy system models, this do-a-thon could further discuss sensible linear or conic approximations/relaxations (or ideas how to approach these).
- populate and re-organise the openmod wiki page on gas network datasets  including a list of relevant modelling components and their attributes and discussion of gas flow formulations
- steps / a todo list to get a simplified European gas network dataset
NB: if there is a big interest, it might be worth splitting this do-a-thon into a “dataset” part and a “modelling” part.