Creating Easily Composable Interfaces for Data Harmonization

[TheCedarPrince]

From @00krishna:

One thing to also think about, is creating a way for users to setup their data in Julia types that allow them to run different types of analyses. So the spatial data is one things, but setting up their data in Graphs.jl format for network analysis, or also loading data to the GeoStats packages for statistical analysis. There is also the idea of loading the data into some more approximately continuous format--meaning something that allows for interpolation between datapoints or density estimation. I have not thought through this part very much yet, so it would need to be refined.

Like one challenge I see with Julia adoption, is that there is a tendency to write stuff for people who already have expertise. For example, if you look at a lot of packages, say even the SciML ODE packages, the docs are written for people who already have a lot of familiarity with ODEs and DAEs, and such. So that can be intimidating for new users, or grad students, etc., who want to try these things but don't have a lot of experience. Hence setting up some of these data into the appropriate data structures will help those folks get over the first hurdle.