Timeline for Forecasting monthly visitor count from daily values
Current License: CC BY-SA 4.0
15 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jan 7, 2021 at 17:17 | comment | added | whuber♦ | That would provide more information than non-overlapping windows and would be more challenging (but potentially more rewarding) to analyze. | |
| Jan 7, 2021 at 15:38 | comment | added | Clock Slave | How do you mean? | |
| Jan 7, 2021 at 15:36 | comment | added | whuber♦ | It could get really interesting in that case if the windows overlap ;-). | |
| Jan 7, 2021 at 15:36 | comment | added | Clock Slave | I can make this a windowed statistic if a suggested method to forecast requires that kind of input. I have login and session info which I use to generate this table | |
| Jan 7, 2021 at 15:34 | comment | added | Clock Slave | Monthly unique vistors: For a particular month, say July, I fetch data of the users that logged in at least once. From that pool, I see the distinct number of users for the month. That becomes monthly unique users. Yes, so for July there will be one value only because that is the truth value. I want to use the number of distinct transacting users for a given day to forecast what the monthly total will be. Thanks! | |
| Jan 7, 2021 at 14:58 | comment | added | whuber♦ | Thank you. What does "monthly unique visitors" mean when it's associated with individual days? Are you saying there's really just one value for the calendar month of July 2020? Or is this possibly a rolling or windowed statistic? (+1, by the way, because the nature of your question is now clear enough and I think it's a new one for this site.) | |
| Jan 7, 2021 at 14:56 | comment | added | Clock Slave | I have added a snapshot and a little description of the data I have. | |
| Jan 7, 2021 at 14:55 | history | edited | Clock Slave | CC BY-SA 4.0 |
added 329 characters in body
|
| Jan 7, 2021 at 14:46 | comment | added | whuber♦ | You need information about repeat visitors. For instance, one person visiting your site each day for six months gives a series $(1,1,\ldots,1)$ of $183$ counts and $183$ distinct people visiting your site can give the same series of counts. Are you saying the monthly counts are of distinct visitors rather than totals? That appears to be so, but offering a small example of your data could help. | |
| Jan 7, 2021 at 14:15 | comment | added | Clock Slave | I have added it in the edit now. Please check. | |
| Jan 7, 2021 at 14:14 | history | edited | Clock Slave | CC BY-SA 4.0 |
added 178 characters in body
|
| Jan 7, 2021 at 14:07 | comment | added | whuber♦ | You haven't described any kind of data that would enable you to do that. | |
| Jan 7, 2021 at 13:57 | comment | added | Clock Slave | I could add that up but that would lead to double counting of users visiting the webpage. I am looking for monthly unique users. | |
| Jan 7, 2021 at 13:25 | comment | added | whuber♦ | Umm...why not sum the numbers of visitors (predicted or actual, depending on what you have) over the days of the month? | |
| Jan 7, 2021 at 4:23 | history | asked | Clock Slave | CC BY-SA 4.0 |