Hi Ken and thanks for answering!
I knew about user devices table, but if I'm not wrong in that table are only stored the devices that have accessed Moodle via the mobile app (a device that only has connected to Moodle using a browser isn't there); anyway, AFAIK there is no way to join that information with the log so I cannot know how many clics from a user came from that mobile device and how many came from a computer.
As a workaround what I've done is this:- 1.- Querying the log I can know the percent of clicks that came from the mobile app (origin='ws') by month:
SELECT to_char(to_timestamp("timecreated"),'YYYY') as year, to_char(to_timestamp("timecreated"),'MM') as month, count(*)
FROM m_logstore_standard_log m
WHERE m.origin='ws'
GROUP BY to_char(to_timestamp("timecreated"),'YYYY'), to_char(to_timestamp("timecreated"),'MM')
ORDER BY to_char(to_timestamp("timecreated"),'YYYY'), to_char(to_timestamp("timecreated"),'MM') ;
- 2.- I can know from our analytic tool the percent of visits that came from mobile devices
Using these percents I can have an approximation of the usage of the mobile app vs. usage of mobile devices.
Anyway, in order to have accurate data I think it should be done one of these improvements:
- Register some info about the device in the Moodle log
- Make it possible to track the app usage via an analytic tool.
Ondo izan.
P.S.: As an extra ball, I've done some changes in the query to know how many distinct users connected to Moodle using the app by month:
SELECT to_char(to_timestamp("timecreated"),'YYYY') as year, to_char(to_timestamp("timecreated"),'MM') as month, count(distinct userid) as distinct_users
FROM mdl_logstore_standard_log m
WHERE m.origin='ws'
GROUP BY to_char(to_timestamp("timecreated"),'YYYY'), to_char(to_timestamp("timecreated"),'MM')
ORDER BY to_char(to_timestamp("timecreated"),'YYYY'), to_char(to_timestamp("timecreated"),'MM');