Does anyone have the data behind the MQ9 response curves? At some point they were created and must exist somewhere in a table…? Of course the points could be taken off the graph, but it would be so much nicer to fit the actual measurement with an exponential and use curves to initialize and produce output from the sensor.

I think the only way is to extract the points from the graph (as you have already mentioned). I found no straightforward data in the intenet. Once I used WebplotDigitizer to extract the points. Then I used microsoft excel to derive the equation. It’s really a lengthy process. I found no alternative. By the way, if you need proteus libraries of MQ-9, you can see here: Gas Sensor Library for Proteus - The Engineering Projects