Given all weekly music charts (of the most popular music tracks in my country) from 1966 till today,
I want to write a program that is able to generate all kind of reports out of that music chart data.
Eg. The list of the xx most popular hits seen over a certain (and why not, the complete) period.
To find the most popular songs, I give each song points based on its position in the charts.
Eg. # points = (# of items in the chart - position)
That way, in a top 50, the track on the first position gets 50 points, while the last track gets 1 point.
However, I have a problem with that.
I discovered that in the sixties I only have charts of 20 tracks.
In May 1970, a switch was made from a top 20 to a top 30.
In the 80's to went to 40 tracks and later on they switched once again to a chart containing 50 tracks.
So, for the period 1966 till today I have weekly charts containing 20, 30, 40 and 50 tracks.
It's clear that with the points system I explained above, the older music can never beat the newer music,
1) a top position in 1966 generates 20 points, while a top track of today gets 50 points
2) in 1966 a track had only 20 positions to score points (number of weeks in the chart), while today that is 50 positions.
Question: Can anyone think of a way to give points to music tracks in a way that eleminates the above mentioned disadvantages?
I know that one solution is to only consider the top 20 tracks of all charts.
But I wonder if it's possible to use all information AND yet keep it "honest".