We're currently using 2017 jobs data, falling back to 2016 for states that don't have 2017 data files (AK and SD). New data has come out, so we want to update the analysis script to use 2018 data, but still fall back to 2016 if it's not available (AK only).
We could manually load the cache bucket with all the 2018 files, but we don't necessarily need to, since the auto-caching should fill it up as they're needed.
We're currently using 2017 jobs data, falling back to 2016 for states that don't have 2017 data files (AK and SD). New data has come out, so we want to update the analysis script to use 2018 data, but still fall back to 2016 if it's not available (AK only).
We could manually load the cache bucket with all the 2018 files, but we don't necessarily need to, since the auto-caching should fill it up as they're needed.