Techsurvivors

Archives => 2008 => Topic started by: Mayo on October 26, 2008, 01:19:55 PM

Title: Is This Possible?
Post by: Mayo on October 26, 2008, 01:19:55 PM
Yesterday I had this idea to create a line-graph of blood test results over time.  The tests are varied so they don't measure the same thing so I am unclear how a graph can be constructed in this case.  I am interested in seeing how the results varied during the past couple of years. The test results are interrelated because a change in one result may be reflected in another test or the results of two or more tests taken together may be more enlightening than just considering individual results. If I can figure out a way to incorporate changes in medications, so much the better... I envision a graph of different color lines that might reveal how test results have been influenced by changes in medications, etc.

I have never done anything like this and I assume that Apple's Numbers program might be used to accomplish this. And I already own iWork...

If graphing isn't an option perhaps a basic spreadsheet is an alternative. Presently I have a small stack of test results which my health care providers and I have to flip through and use the data to connect the dots between various tests and changes in medications and supplements.
Title: Is This Possible?
Post by: Gregg on October 26, 2008, 02:00:36 PM
QUOTE(Mayo @ Oct 26 2008, 01:19 PM) <{POST_SNAPBACK}>
The tests are varied so they don't measure the same thing so I am unclear how a graph can be constructed in this case.


You'd just end up with separate graphs, wouldn't you?

I don't know about the graphing capabilities of Numbers. But I have the Missing Manual for Office, and I'm quite sure Excel could generate graphs from your data.

Good luck! smile.gif
Title: Is This Possible?
Post by: Xairbusdriver on October 26, 2008, 03:52:32 PM
The easy part would be using Numbers. Just select the "blank" page template and it will open with columns for A through M and rows 1 through 45. There are no numbers in there, so I guess that's why they call it "blank!" dntknw.gif
More columns and/or rows is simple, also.

The hard part, you have already mentioned. How should the data be compared, in the first place. And how should the meds be related? It actually sounds like a good use for FileMaker. Enter all the basic info in records. Then sort on no meds, one med, two meds, etc. A graph from each set of records might be meaningful. But the fact that the values are not always exactly comparable makes interpreting the resulting graphs rather imprecise, IMHO.

Sorting may also be possible in Numbers, I've never used it.
Title: Is This Possible?
Post by: Gregg on October 26, 2008, 04:11:27 PM
Sorting is definitely possible in Excel. I kinda doubt Numbers does that for you.
Title: Is This Possible?
Post by: Mayo on October 26, 2008, 11:26:43 PM
I'll start by entering the data on a spreadsheet; that will at least make it easier to compare numbers over time. I don't know why I didn't think of this before now... A couple of days ago I noticed that my wife had created an analog spreadsheet detailing her test results. D'Oh!
Title: Is This Possible?
Post by: Xairbusdriver on October 27, 2008, 02:11:48 PM
IMHO, 'sorting' is not the problem. It's getting that collection of data into a usable form, in this case a graph. That's why the "Find" combined with a 'multi element sort' sounds needed. You don't really want a graph showing values from widely different types of tests, that could give wildly distorted results/impressions. It sounds like the basis of any comparison graph is to use one key variable, like the type of test, and build the graphs with the other variables and values with them.

That's why FMP is so easy to use, no need to keep track of what row, column you are in. An easy to use/view 'form/layout' is a snap! All finding/sorting/presentation is fully scriptable or can be manually done, if desired. You might be able to use Bento, but it is much less capable (than FMP) when it comes to mathematical formulae.