Data scientists often deal with having too little data, but how do we deal with having too much data?
Your engineer asks you, “how much can we subsample this data without error?” The answer is, of course, is not at all. And yet... what if we are willing to get wild and take on error? How much faster can we go?
Join Philip Geurin, a data scientist at Uber, as he sifts thousands upon thousands of data points per second... all to predict a tiny number.