Needs Thought/Ideas - Data Compilation/Aggregation/Output

[scrupul0us]

I've given myself a little challenge and to be honest I'm a bit lost as to how to implement it

Basic Idea:
-For two hours a week monitor the stream stats of a web radio website
--grab the number of users listening on 64K and 128K
-Compile these stats at the end of the hour, store them and then be able to use them to display other stats
--Peek users, avg users, total users, user density (break out for each 15 minute span), peek period (e.g the most users were during the second quarter of the first hour)... and maybe a few more I haven't decided

Now... what I AM able to accomplish so far:
-When exactly to start collecting stats (i can ereg the page and detect that its ok to start)
-How to grab the current number of listeners per bit channel

What I'm struggling with is the actual "recording" of the data

I think I want to poll the website every minute, this way I get 120 data points for the duration which should reflect some pretty accurate results

At this point I'm open to suggestions on how to even begin this let alone code it... so please feel free to express your thoughts and ideas

Thanks!

Brian

[leatherback]

Hey Brian,

Wouldn't your webserver access log give you all the info? You can then use a log analyzer to create the stats.

That being said: I would just store the info in a database, and at specific times (e.g., once a week) run a buynch of queries to create the stats, and clear out the log in the database, of you do not need the details

[scrupul0us]

this is off a remote server, otherwise i would agree

i also agree with using a database to store the data... my issue is how to get the "program" running...

will i have to setup a cron job?

heres what im faced with... almost every tuesday night my buddy does a radio show... i record it and post it for him... he had asked me if i could compile a page with stats for the show etc... now, some times he doesnt do his show due to work (hes in the military)... that being said id need something to flow like this

1) recognize its 6:58pm EST and start checking the website to see if he's playing
--he may com in at exactly 7pm or be running late and be online by 7:05 or so

2) if hes online start recording stats every minute for 2 hours and continue to check that he still is online

3) if hes off line @ lets say 7:10 then he probably is going to be on and we should just dump the routine

4) once hes off line stop recording stats