- Get full technological class in redbook database (incomplete parsing) - Clean and integrate into pipe the data collection/parsing procedure - Rewrite R-friendly technological matrix construction (mongo -> sparse R matrix instead of old csv -> sparse R matrix)