1

I have a dataframe with each row representing a visit by a user, including a userid column and a date column. I'd like to calculate the number of unique visitors to date (not just unique visitors for that date - unique visitors from the first date of dataframe until the row date) for each date in the dataframe

Dataframe:

date           clientid
2015-01-01     12345
2015-01-01     12346
2015-01-02     12345
2015-01-02     12347
2015-01-03     12347

Expected output:

date         unique_visitors_to_date
2015-01-01   2
2015-01-02   3
2015-01-03   3

This code would work in theory but the for loop is stuck in purgatory and also I think for loops are not the "right" way to do it.

for i in range(0,519585):
     visit['cumulative'].loc[i] = visit[visit.date <=
 visit.date.loc[i]].clientid.nunique()
cs95
  • 379,657
  • 97
  • 704
  • 746
Ali
  • 11
  • 1
  • `df.groupby('date')['clientid'].nunique().reset_index()` – cs95 Jun 27 '19 at 21:44
  • Hi Ali, I think your problem is a bit more than the marked dup, I hope it will be reopened but in the meantime, try this `df.drop_duplicates('clientid').groupby('date')['clientid'].nunique().cumsum().reindex(a.date.unique()).ffill().reset_index(name='unique_visitors_to_date')` – Ben.T Jun 27 '19 at 22:08

0 Answers0