Poisson Regression - Roundup

June 24, 2018
Poisson Regression on data provided by Circle K
This experiment aims to use the data sets provided by Circle K Hong Kong. It grouped the date and time of the transaction in the week of day (i.e. 0-6) and hour (i.e. 0-23). After that, by using group data, we select the data with the same categories, same store type as well as same day and hour as one set. And then try to run the Poisson Regression to predict the quantity of sales during a particular day and particular time in a region. The result is satisfied by having around 0.5 coefficient of determination.