good correlation doesn't imply causation

there are several sites with good examples

these last two are from  tylervigen.com

scatter plus a correlation coefficient

sales scatter graphs

I am unaware of the person(s) who came up with this idea for scatter graphs and it's hard to find this information out because several resources have used the idea

(I'd be grateful if someone could let me know)

the commodities and their associated scatter graphs usually make sense to students in terms of the explanatory variable

the (product-moment) correlation coefficients, using Excel, in the order presented are:

1.    0.94
2.    0.87 (negative)
3.    0.22
4.    0.31

scatter and correlation

although there is value in plotting scatter graphs by hand, there is also value in using a spreadsheet to automatically draw a line of best fit and give its equation

people argue, generally rightly, that ICT can free up time to consider interpretations, evaluations and other analyses

using ICT, questions can be reversed: students could be asked to generate data sets with particular correlations

students could be asked to guess these,
following an understanding that the coefficient varies between -1 and +1

some scatter graphs with their correlations