h2g2 The Hitchhiker's Guide to the Galaxy: Earth Edition

Here is some analysis of the figure for the number of people online at each sample:-

count| avg| stdev
-----+---------------+-------------
92|56.554347826087|10.2928163784
(1 row)

count shows the number of reading made.
avg shows average number of people online.
and stdev shows the population standard deviation of the data.

Typing this into my calculator to find error bounds for a 95% confidence interval I get:-

56.5545 +/- 2.1315.

Analysis of the other data to follow.

I'm user 87448 and I'm online all day at work - Category A. I'm usually only actively online for several bursts: once in the morning to catch up on backlog, once at lunch, once near the end of the day. I sometimes check for activity inbetween. Probably adds up to 2-4 hours. I am also user 148944, but only log on as that user roughly every other day for about an hour.

I've created an article to house the current statistics, it can be found at http://www.bbc.co.uk/h2g2/guide/A672554. I'll try and update it daily/weekly.

U55660 reporting in

I'm normally on the lurking list for around 5 - 6 hours a day, although if I'm working from home I may appear to be here for around 16 hours. (ie all the time my PC is on )

My actual active time on Hootoo is nearer 2 - 3 hours a day

My primary U number is 99500, and depending on what time of the year it is, I'm resting on a razor-thin edge between A and B. Makes me realize just how much of my life I've wasted on this site.

I think the "waste" part has been waiting for the servers, as d'E mentioned...

The statistics so far seem to be suggesting that the Gurus are a bunch of s who need to get lives.

Hmmm.

Amy the (Guru) Ant

Amy: the stats for the number of hours are for the /average/ research, it is only indirectly affected by whether the group are s or not.

s/research/researcher

I've updated the stats at http://www.bbc.co.uk/h2g2/guide/A672554. Plus I've put up a graph linked to from that page.

Interesting. Only, could you possibly improve the layout, and reduce the number of significant figures as it is hard to read.
Munchkin, always too lazy to use tables.

Yeah, I know but the output is directly copied off some hacked together sql scripts tied together with a large pile of bash and perl scripts.

Munchkin: the page is now up with everything cast to 4byte floats not 8byte float so everything is to 6 sf.

I don't understand anything!


Is there a key to what the columns mean somewhere?

Here's the key as they are a bit cryptic:-

First table:
listname - I have multiple lists set up so that unlike the first project I can compare results obtained using different groups of researchers. The original group used the group named 'agb' which contains researchers who have met Archangel Galaxy Babe in real life.
count - the number of samples collected for each list
avg - the average value of p for that list - see the orignal page for what p is
stdev - the population standard deviation of the dataset

Second table:
count - total number of samples of r
avg - average r
stdev - population standard deviation

Third table shows the consolidated final results in columns 'total' and 'hours':
listname - see the first table above
friends - the average number of these people online at once
active - the average number of researchers online at once
count - the number of people on the list
total - a (highly inaccurate) estimate of the total number of researchers who are considered 'active'
hours - using 'total' an estimate of the number of hours the 'average' reseacher spends here. It varies a lot depending on the usage characteristics of the group.

In my mind I like the data obtained from the nhm list best as the people/animals/things featured at the natural history museum contain a broad spectrum of people who live, eat and sleep and here and those who pop in every so often.

Hope that clarifies things.

Forgot to add another reason I like the nhm data:-
the first table shows the stdev for 'nhm' to be the lowest, so that means that its nearly the same number of people online from this lists, hopefully indicating that the list contains people from a wide range of timezones and times when they can connect.

That last point of your is a very good thing to note x

Thanks, I will do more through statistic analysis of the data sometime at the end of January, when exams are gone and I have oodles of data to play with. (Plus with luck an upgraded copy of Postgres with better support for views.)

Yep, thats a bit easier to read, ta.

looks good. I'll just point out there is a minor typo on your fraph page... Te y-axis

I can't see any problem apart from weird ticks on the x-axis and the fact that the y-axis does't cross the x at 0. I've fixed the second problem, so when I update tomorrow morning it'll be fixed. Although I think I might try rrdtool to do the graphing as I will end up with globs of data and I don't want to strain poor wee gnuplot.