statistics over pairs of values

[blum22]

Hello community,

I am working on pairs of values generated by some data analysis. They come in two 1D-arrays of equal sizes (typically thousands of values) that I then cluster to plot in an XY graph, much like described here.

However, my analysis is such that I can get a lot of pairs for a given X-value, and my X-values are integers. Therefore, I would like to perform some basic statistics on each set of XY-pairs for a given X (number of pairs, mean, median,...), but I don't see any obvious way of doing that the way my pairs are currently organized.

Am I missing something here, or do I have to sort all my pairs by X-values?

The screenshot is an example of my XY graph.

[santo_13]

Attaching sample data and any related VI will help us to play with our idea to solve your requirement.

[FireFist-Redhawk · Accepted solution]

It seems to me that yes, you would have to separate your Y values into subarrays of Ys that are associated with each X. See if this skeleton code helps at all.

Spoiler

[altenbach]

It would help to see some typical data. For example are the pairs already sorted by X or not?

I would use a map with x as key and an array of Ys as values. Once the map contains all data, iterate over all keys and do the statistics on the y (array size (=# of parts, mean, etc.)

---

@blum22 wrote:

... perform some basic statistics on each set of XY-pairs for a given X (number of pairs, mean, median,...),

---

Can you give a complete list of the needed statistics? Some can be done ptbypt, some not.

[blum22]

Thanks! That's indeed what I had in mind, but cleaner than how I would have done it.

Works great.

[blum22]

Thanks a lot for your feedback. No, the pairs aren't sorted as they are generated without clear order, but I will keep maps in my mind for future use (I didn't know they were introduced in LabVIEW).

[altenbach]

Here's how you could use a map.