Upgrade Task:
OJs 2.3.3.3-3 towards OJS2.4.8-2
After following instructions listed here:
https://pkp.sfu.ca/wiki/index.php?title=PKP_Statistics_Framework#Summary
The result is as follows:
Why is the output data changing from pre-upgrade to post-upgrade? Could this be new filters applied to the old data that was acquired - such as bots been filtered out of the older data?
Hi @tretief,
The views report in older versions (previous the introduction of the new stats framework) was not logging any information about the user that accessed the content. So there is no way to get that data, migrate it and filter double clicks or bots, because there is not enough information for that.
The purpose of the views report was really to give an estimate on the usage of each content, rather than an exact number. That’s why OJS also had other stats plugins, like counter and timed views.
Now your post upgrade numbers are really strange, compared to the pre upgrade. I see an increase, almost doubling the number of PDF and HTML galleys views, and a big decrease in EPUB.
One thing that affects the migration and will affect the post upgrade numbers is database consistency. I mean, all data related to the galley (article, file) must be present. If, for some reason, there is only a galley record, but it’s associated article or file records are not in the database, this galley record (the one that has the view count) will be ignored when migrating.
This could explain the decrease, but not the increase in numbers. I would have to take a closer look to your database to try to figure this out. I have limited available time, but I can try, if you don’t bother sending me a dump.
Cheers,
Bruno
How can I send you the dump?
