Sharing pre-calculated covariates for initial model fitting?

[sigmafelix]

I understood that we will deploy a R package, downloaded data, and pipeline scripts (if not included in the package). Since covariate calculation for fitting the base learners in 2018-2022 take a significant amount of time, I think the pre-calculated covariates are also needed to be shared here or elsewhere. A packet of pre-calculated covariates may include:

• Sites: Spatiotemporal covariates and daily PM_2.5 measurements as a data.frame or data.table (in csv/rds format)
• One kilometer spatial resolution grid points: for reproducibility, it will be good to have a fixed set of grid points in EPSG:5070 CRS

The pipeline code will update the existing 2018-2022 measurements and covariates with those at new time points (e.g., Q1 2024, Q2 2024, ...). By doing this, users will not need to download terabytes of base datasets from the internet for five years.