Counting Observations

[travisweiland@gmail.com]

Is there a way to easily count the number of points in a region of a graph? For example if I want to find the Quadrant count ratio would there be an easy way to count all the points above the mean of each each variable in the graph? Similarly if I added a function f(x)=x is there a way to easily count the point above and below the line to compare?

[Bill Finzer]

Hi Travis,

I’m afraid all I can come up with is rather crude. It’s shown in the attached screenshot.

There are two scatterplots, one of which has the two plotted entities that define the quadrants.

The second scatterplot is of the same attributes. It is showing the count of the selected points. This option is in the eyeball menu—Show Measures for Selection.

Probably not what you want. Maybe someone else has a more elegant solution?

Bill

Attachments:

1. 

   Counting-Selection.jpg
   

[Dan Damelin]

This doc shows some options. One can calculate a count or calculate a quadrant and then color by quadrant. If you turn on “show measures for selection” in the graph, then clicking on one of the quadrants in the legend will select just those and show a count in the upper right. https://codap.concord.org/releases/latest/static/dg/en/cert/index.html#shared=https%3A%2F%2Fcfm-shared.concord.org%2FhsqIEZxkwpcYCUHSyR35%2Ffile.json

[Dan Damelin]

Bill if movable values could be captured in the formulas, then one could manipulate the line and see reflections in the formulas automatically updated. Not sure how often a feature like that would be used though.

[travisweiland@gmail.com]

That is helpful to know.

A super useful feature would be if the count measure could count the observations on either side of a function or value. In research on finding a line of best fit one strategy kids often use is getting the same number of observations above and below a line. Also the quadrant count ratio is a development measure building towards R that is suggested in the NCTM essential understandings books and GAISE II framework so I could see it being helpful there to. In small datasets we can often eyeball it but in moving toward more of a data science big data set focus it becomes challenging so a technological tool would be helpful.

Thank you all for all you do!

[lperez2@wested.org]

Picking this thread back up as I found myself trying to find a way to do the QCR as well this morning. I will try Dan’s solution but I do wonder if this measure would help us in the Data Fluency.

[Dan Damelin]

For univariate distributions you can add a moveable value and see counts (and/or percentages) of data points on either side of the value. See screencast.

• This reply was modified 1 year, 7 months ago by Dan Damelin.

[Andee Rubin]

Just want to chime in here in support of an easy way to calculate QCR.  I discovered this measure recently and find it a very useful stepping-stone to understanding correlation coefficient.

[esatuavci@gmail.com]

Hi,

Data that has the same value as the mean in the Quadrant Count Ratio which region should be included? So, the axis passes through the middle of the data.

 

[Author]

Posts