Practical Predictive Analytics

Identifying members with multiple purchases

Try this query on your own in a new script to determine which members had multiple purchases. There should be 89 of them. In SQL query parlance, the multiple purchases clause is indicated in the query by count(*) > 1:

dups <- sqldf("select member.memberid,count(*) from  member left outer join  purchases_filtered on  member.memberid=purchases_filtered.memberid2  group by member.memberid having count(*) > 1")