Hi, I'm also thinking about taxonomies. Not event level data, but most common case as data product is a big phat table (>1000 columns from which we make selections) with with columns that have sometimes unique values per row, but are often less granular.
Currently we are very user focused and have short prefixes (abrrevations) for 'aspects' of the data and sometimes at the end something like "_code". So very user oriented and hybrid (not many strict recipies) for ourselves sometimes confusing.
I would like to have a first prefix based on on granularity (since we are mixing so many levels into one table), than aspect (if still needed) and at the end always a clear indicator what sort of variable it is (ID, CODE, DESC(RIPTION)). To make it clear for ourselves and do the engineering / ml stuff we want to do easily.
Anyway, so far my own ramblings. I wanted to write I was interested in the template ;-)
Voldermort-like concept of ‘self-serve analytics’
^^^ made me lol, thank you
Hi, I'm also thinking about taxonomies. Not event level data, but most common case as data product is a big phat table (>1000 columns from which we make selections) with with columns that have sometimes unique values per row, but are often less granular.
Currently we are very user focused and have short prefixes (abrrevations) for 'aspects' of the data and sometimes at the end something like "_code". So very user oriented and hybrid (not many strict recipies) for ourselves sometimes confusing.
I would like to have a first prefix based on on granularity (since we are mixing so many levels into one table), than aspect (if still needed) and at the end always a clear indicator what sort of variable it is (ID, CODE, DESC(RIPTION)). To make it clear for ourselves and do the engineering / ml stuff we want to do easily.
Anyway, so far my own ramblings. I wanted to write I was interested in the template ;-)