Presenting Mixed Data Series on One Scale
From: John Gregory (jaygreg90_at_hotmail.com)
Date: 07/14/04
- Next message: Stefano Ghirlanda: "Re: Presenting Mixed Data Series on One Scale"
- Previous message: George Avery: "Re: Linear Regression"
- Next in thread: Stefano Ghirlanda: "Re: Presenting Mixed Data Series on One Scale"
- Reply: Stefano Ghirlanda: "Re: Presenting Mixed Data Series on One Scale"
- Reply: Richard Ulrich: "Re: Presenting Mixed Data Series on One Scale"
- Messages sorted by: [ date ] [ thread ]
Date: Wed, 14 Jul 2004 12:37:21 GMT
A group of folks interested in understanding economic business cycles for
purpose of investing are trying to find a way to graphically present mixed
data on one chart. If multiple scales ("Y" axis) are used, the graph becomes
messy... too busy.
Patterns are more important to the group than concrete values. They're
lookin' for trends and relationships.
There's common agreement that the data should be "normalized" (don't even
know if that's a legitimate term in this case) by taking the highest number
of the series that has the highest value and divide it into the highest
value of the series that has the lowest values in order to establish a
proportion. Then, using that proportion, divide it into every date point in
the series having the highest values. When done, move to the next highest
series and repeat the process.
An alterative but very similar method proposed was to ignore the high values
and simply establish ratios for the same year across all series then divide
the respective results into each of the data points within their respective
series. You end up with a different set of numbers but surprisingly (to me)
you end up with the same relationship or proportion each number has within a
series to its neighbors. Thus, the same "pattern" is projected and all will
fit on the same scale.
Comments please? Anyone have research background who has dealt with this
before? We may be reinventing a wheel here and going to a lot of trouble for
nothing.
- Next message: Stefano Ghirlanda: "Re: Presenting Mixed Data Series on One Scale"
- Previous message: George Avery: "Re: Linear Regression"
- Next in thread: Stefano Ghirlanda: "Re: Presenting Mixed Data Series on One Scale"
- Reply: Stefano Ghirlanda: "Re: Presenting Mixed Data Series on One Scale"
- Reply: Richard Ulrich: "Re: Presenting Mixed Data Series on One Scale"
- Messages sorted by: [ date ] [ thread ]
Relevant Pages
|
|