prepare dummy data

[Steve Harris]

explore likely data in a jupyter notebook
then use faker or similar to generate similar dummy data
that can then be used for teaching/working
but is appropriately aligned to the original source query

[Steve Harris]

had a go and shared as vignette
sampled from the empirical distributions
so will be realistic wrt to univariate characteristics but all correlations broken
should do for testing

[Steve Harris]

done
see https://hylode.github.io/HyStakes/vignettes/prepare-synthetic-data.html#prepare-dummy-data-frame and https://github.com/HYLODE/HyStakes/blob/main/data/dummy.db