in my last 10 years of generic stem research/data analysis, I've gravitated towards adopting "standard" data formats, e.g. generic timeseries in hdf5 and "small

in my last 10 years of generic stem research/data analysis, I've gravitated towards adopting "standard" data formats, e.g. generic timeseries in hdf5 and "small" utilities that do simple things to the data and produce plots. this allows me to extend/adapt/remix my pipelines much faster than I could by developing either custom scripts for each case or one big app to solve everything. of course I didn't invent this, it's pretty much the UNIX way for non-ascii/structured data.