Web Site Page Visit Stats (Relative Popularity)

Greg

New Member
#1
I've started to create a statistical analysis of relative popularity for visits to the articles on a web site.

I'm not a statistician, but have made what I feel was a pretty good attempt at the problem.

I have a table with one article listed per row where I want to display a bar chart (0% - 100%) of each article's relative popularity to the others.

The information I have to work with:
* Date article was published.
* Number of visits article has received.

From these two bits of data I have come up with the following formula, which seems to work okay (with a few issues):

p = (v / a) / MAX(v / a) * 100

where:
* p = Relative Popularity (p)
* v = Total Visits (v)
* a = Article Age in Days


This seems to do fairly well at placing giving each article a popularity rating between 0 and 100. But I have the following problems:

1. If the article is brand new, a = 0, thus causing divide by zero errors.
2. If an article is new and has a few visits, it might rank as the most popular article on the site compared to one that's been there longer with many more visits. So the relativity factor isn't quite right.

I need a statistician to help out from here. Any ideas what I can do to correct and improve this formula?