Statistics question - how to do this without screwing up my stats?

AARRRGH

Hello all

I hope someone can help me with this.

I am implementing a job statistics function on one of my websites. In a nutshell, I monitor the number of job adverts on various jobsites. I want to create a historical line chart so people can see if there are more or less jobs than last month, etc.

This is my problem:

Lets say I am monitoring the job adverts on site1, site2 and site3. In 6 months time I then add site4. Obviously this will greatly increase the number of jobs available so will screw up my statistics.

Is there any way I could add site4 and somehow merge the data into my existing data without screwing up my previous statistics?

Any advice appreciated.

Thanks!

RoundTower

I don't think there is any sensible way of doing this until you get a few months of data from site 4.

CJC86

I would have thought that it would be sensible to work with the 3 sites separately since there is probably a reasonable overlap between the 3 sites anyway (hence, at the moment you may be double-counting some jobs). If you deal with them this way, then obviously you could add in a fourth site, and just have less historical data for that site.

For a "Total number of jobs" graph, perhaps it would be better to leave out site 4 until you have collected a few months data, as RoundTower suggested.

Thoie

It won't be at all accurate, but you could put the average of Site1, 2 and 3 as the missing stats for Site 4.

The other, more accurate way might be to show the average number of jobs across the sites, month by month, rather than the total - see attached jpeg.