- CSV files from https://www.data.gouv.fr/fr/datasets/agribalyse
- alternatively from https://nexus.openlca.org/database/Agribalyse
Note: the files from data.gouv.fr are the files from the Nexus OpenLCA converted to a more easily readable format, i.e. csv files, according to data.gouv.fr
Tables are explained at https://www.data.gouv.fr/fr/datasets/agribalyse More info from the methodology page https://doc.agribalyse.fr/documentation/?cacheBust=1634652873522
Note: according to this person, the table TBL_EXCHANGES.csv
contains 3.5 million lines (one line, for one technosphere or environmental exchange). However, 75% of these are equal to 0 and useless to parse. Saves a lot of time to just filter them out before writing the database to brighway! (discovered at own expenses)
Inspired by Massimo's import for Exiobase.
- Load all the CSV files as Pandas [done]
- Identify all activities & products and create a dictionnary [done]
Issue here: the matrix is not square, there are more products than activities (missing activities can be replaced by dummy activities, hopefully)
- Write the dictionnary as bw2 database [done]
- Create a correspondance table between bw2's biosphere3 & the biosphere from Agribalyse [partly done, requires help/improvement]
Issue here: not so easy to get a 1-to-1 correspondance between the two, neither automatically, nor manually
- Loop on all biosphere & technosphere exchanges, and save exchanges to the database [done]
Issue here: need speed improvement, need handling of issues (missing activities/products)
- Do LCA calculations for all products/processes & validate them against implementation in SimaPro or openLCA [not done]
Validation is THE important step!