Subscribe Bookmark RSS Feed

Align the values on graph to start at the same point?

prmiszkiewicz

Community Trekker

Joined:

Aug 7, 2014

I have some data from measurements and I need to compare the results between the several tests (graph of value over time with test number as legend). My problem is that different tests have starting values slightly different (later shape of graph is similar, but they are e.g. slightly above or below because of that start difference). Is it possible to 'calibrate' the graph so that the start value is always a fixed one, while the later development of values is just like original (after applying that difference to make start equal)?

I know I can manually calculate the differences and apply an Excel function to do it, but I'm looking for a faster and less time-consuming way.

2 REPLIES
stephen_pearson

Community Trekker

Joined:

Oct 6, 2014

If the results from each group of tests have a common identifier name then the join command is the easiest way to combine tests.

If the time between groups of results is big relative to the time between results then cluster analysis could be used on the date/time (you would need to stack all the results so they share a single date/time column).

If the data shows long term trends with time you can use fit Y-by-X on each date/time test pair. Fit a spline to each graph and then save the spline equation. The columns can be combined in a new data table and each formula pointed at a common date/time column.

Byron_JMP

Staff

Joined:

Apr 26, 2012

I'm sure that there are at least a dozen ways to solve this.

In the attached table there is a formula that resets an Aligned Time Column each time the run/batch identifier changes. In this approach I subtracted t=0 from the time stamp of each run.7731_figure 1.png