I'm not sure if this pertains 100% to the question, but here's a blog post that involves filtering in a chart. I believe you can apply this to a scatter plot with some additional work. The key is that the graph chart is using ExecuteSQL to find the results based on values in global fields. Are you feeling good about using ExecuteSQL?
Adding to Jeremy's response:
You don't want to filter the graph, you want to filter the data it's based on.
In correspondence with the data sources that are offered for an FM chart, that would mean:
• change the found set (i.e. perform a Find)
• change a/the relationship
• retrieve different sets of delimited data
Which one you should choose depends not the least on your skill level and comfort zone(s) – while ExecuteSQL() usually looks quite promising for novices coming from such a background, one should not forget that this functions still acts within the FileMaker framework …
PS: it may be better to use the official term Chart; when people say 'Graph', they usually refer to the Relationship Graph.
Ack. I used the word "graph" too!☺ I'll edit it out.
Your points are easy to work with. Once I got competent with ExecuteSQL, I started ignoring the other possibilities of the chart and only do the found set/delimited data.
Thank you both. I just created a new layout with the search fields that I want to use and placed the chart right below it. Seems to work okay.
I have another question(s) that I think would be a little bit more difficult: Is there a way to insert a best fit line/linear regression line for the found set on the chart? Also is there a way to determine the correlation coefficient for the found set?