| Author |
Message |
tanneffel
Moderator
|
# Posted: 26 Jun 2013 19:46 · Edited by: tanneffel
Reply
Based on Login records for the past month, here are some statistics on when OTF gets the most logins (click to enlarge):
There is not much variation between the days of the week. Thursday and Wednesday get the most logins, Friday gets the least. Weekends are average.
And logins by hour (click to enlarge):
It would appear that Europeans log on when they arrive home from work, and North Americans do so when they get up in the morning. The North American evening is quieter once Europeans go to bed, and after midnight PST, people are no longer logging in, mostly just chatting until they log out for bed.
Feel free to draw your own conclusions, write them here. Remember that these are login times, not necessarily how many people are on the personnel list at these hours.
|
soka
Member
|
# Posted: 26 Jun 2013 20:23
Reply
Interesting!
How many weeks have gone into those statistics?
the first diagram seems pretty constant - sure there are a few less on Friday, but not *that* huge a difference.
Are those Login-times different on Weekends when some people have a free day and can sleep in / come on at different times?
|
tanneffel
Moderator
|
# Posted: 26 Jun 2013 20:42 · Edited by: tanneffel
Reply
The statistics cover 5 weeks plus a day, and are adjusted in the first chart for the fact that I used six Tuesdays instead of five. The hourly could be slightly skewed toward the extra Tuesday though.
I too am curious how the login times play out over a week's period. I suppose that'll be the next chart.
I was surprised by how constant the days of the week were, and how it was actually midweek and not weekends that people favoured most.
|
mohdanni
Member
|
# Posted: 26 Jun 2013 20:44
Reply
Is each log-in a distinct user?
|
tanneffel
Moderator
|
# Posted: 26 Jun 2013 21:19
Reply
No. It's total logins. So if someone logged in 5 times in an hour to change their name or correct their tagline, each one was counted.
Though since time-outs no longer happen, people like Hydra push the number of logins down whenever he sleeps behind Guinan's bar.
|
soka
Member
|
# Posted: 26 Jun 2013 21:28 · Edited by: soka
Reply
But after he wakes up he usually doesn't talk before he leaves, so he doesn't push it down after all.
|
trinity
Member
|
# Posted: 27 Jun 2013 00:12
Reply
Is there a way we can start a longitudinal study to se how this changes throughout the seasons? 
|
tanneffel
Moderator
|
# Posted: 27 Jun 2013 01:21
Reply
*L* Trinity, I'd love if we could do that. We'd have to access the last 25,000 - 30,000 logins most likely.
|
david1
Member
|
# Posted: 27 Jun 2013 04:16
Reply
Depending how logs are stored making a statistics page for the gathered information is simple.
|
hydralisk
Member
|
# Posted: 27 Jun 2013 04:28 · Edited by: hydralisk
Reply
I agree, having access to this kind of data could be very useful for all sorts of applications. Part of solving a problem relies on data like this, rather than anecdote. If it is indeed the case that it's not so much that we're completely empty, but that during the summer people show up at different times, then it would be useful to know that!
|
tanneffel
Moderator
|
# Posted: 28 Jun 2013 03:46
Reply
Adjusted slightly for the day Mela logged in 34 times during 16:00.
Remind me to ask my husband exactly what data and statistics are available to Admirals.
And that is kinda why I started this, Hydra. Solving problems like, why isn't Trivia getting anyone coming out? Saturday at noon always worked for us, didn't it? I'll see if we can go back more than 2000 login records, otherwise it'll be a yearlong project to find out if summer is indeed slower.
I remember back in the day, the Login Records page only covered a few days. Now it covers over a month.
|
tanneffel
Moderator
|
# Posted: 3 Jul 2013 04:24
Reply
HERE IS MY LATEST STATISTIC:
Updates Board Posts from August 20th 2011 to August 19th 2012: 16
Updates Board Posts from August 20th 2012 to present: 136
That's not even an entire year yet. This is an 750% increase in updates board activity over the previous year, so far. I haven't done the breakdown yet, but this is posts from 20 different people too. Some just posted once, but that's still 10-15 people who have posted regularly, leaders sharing their news with the community, inviting the community to re-engage in activities.
Thanks for coming back you guys. I give you all a hearty pat on the back.
|
rpmobsession
Member
|
# Posted: 3 Jul 2013 16:57
Reply
To get more accurate data, couldn't you adjust the table parameters? i'm not sure how you're doing your chart, Tanya, but it seems considering the login records page is set up in a nifty table including username....You could export that information and sort it to delete/remove the same username within a certain time frame (maybe 2 hours?) I know I could do that in excel at least.....but I'm so not a coder/programmer so i'm sure there's a far easier solution.
Maybe someone could poke Ohm and see if there is a public access link that would be able to show updated graphs for people? ...Just a thought.
|
tanneffel
Moderator
|
# Posted: 3 Jul 2013 21:02
Reply
I don't even have excel, and it's still a lot of data to sift through finding multiple logins. The way I did it, editplus automatically counted all logins for each hour and day. I just had to ask it 24 times, then 7. I thought it a rather easy solution, just not entirely accurate if double+ logins are an issue.
|
rpmobsession
Member
|
# Posted: 4 Jul 2013 15:15
Reply
And you can't modify the parameters to search column A for example....for duplicates and then if duplicate is found....search column B (time) and throw out/delete any that are fewer than fewer than two hours? ....I don't know eedit plus but seems most table/data/graphing programs would have such ability...IR's fairly basic table/matrices rules. Was just a suggestion either way. *shrugs*
|
david1
Member
|
# Posted: 5 Jul 2013 19:18
Reply
New toy for you tanya!!
Outpost Chatrooms Stats
You can also adjust it, anything up to 2000, its based on the chatroom login page..
http://xgermz.com/xz/stats.php?a=NUMBER
e.g.
100
500
David.
|
david1
Member
|
# Posted: 5 Jul 2013 19:24
Reply
Ill add a week view tommorow......... over the weekend.... ill be busy this weekend 
|
david1
Member
|
# Posted: 5 Jul 2013 19:59
Reply
Scratch that, ive added in weeks, lol
the week stats is very surprising..... lol.....
|
david1
Member
|
# Posted: 5 Jul 2013 19:59
Reply
our least popular day is the day we hold trivia's on? LMFAO.
|
jamesgilfoyle
Member
|
# Posted: 5 Jul 2013 20:28 · Edited by: jamesgilfoyle
Reply
Gorgeous toy!!
@Trivia Day: That sounds right to me... Should be using tools like Trivia and other events to try to drive traffic on slower days. 
|
tanneffel
Moderator
|
# Posted: 5 Jul 2013 20:32
Reply
Very nice, David. Perhaps we need to rethink Trivia times.
For me, I'd prefer if it was at 16:00. I could get up, do trivia over breakfast, then have my day. At 20:00 it means you can't do anything that day, just bum around home.
|
david1
Member
|
# Posted: 6 Jul 2013 09:08
Reply
Ill fix the bug with the week view as soon as.....
Its using all the data when it actually needs to use whole weeks, rarther then half weeks giving certain days extra login counts
|
david1
Member
|
# Posted: 7 Jul 2013 12:15
Reply
i tweaked it to be more Pretty.
Tweaked the week view, however a new bug has appeared.... basically it chooses whole weeks now, but the current day (i.e. "today") distorts the current days view.
Might tweak to ignore the current week, and just go off previous full week data....
|
david1
Member
|
# Posted: 7 Jul 2013 17:17
Reply
Also, (note to self) modify the scales so they go up in ones.... 0.25 of a user is a little dusturbing...
|
jamesgilfoyle
Member
|
# Posted: 7 Jul 2013 21:17
Reply
David. It's not pretty. It's beautiful!
|
david1
Member
|
# Posted: 12 Jul 2013 09:19
Reply
there we go, the final tweak.
Updated week view to be accurate. Also shows comparison for previous 3 weeks!
|
