Ask HN: As a data scientist/engineer/analyst, what pisses you off in your job?

collecting "data" doesnt mean you can answer the questions you want to know. Hourly GPS data is mostly useless for answering pretty much anything except "where was the car an hour ago", but 0.2second GPS you can do quite a lot. It just always seems like the hourly option gets chosen because it cheaper and easier, and you are now collecting "big data". That option is chosen at the tender stage long before any data people get involved. And it turns out that the location of the car an hour ago is never the problem I get asked to solve with the data.

s/GPS/any other data points/

OP wrote: > people think that machine learning will solve all business problems, when in reality just some simple sql and sorting rules can solve many many problems

Which sorting rules have you found useful? Thanks.

Nothing. I enjoy it very much :)