Currently (Qlik June 2020) Qlik Sense applies an algorithm to compress data if you have more than 1000 data points. This is explained here: https://help.qlik.com/en-US/sense/June2020/Subsystems/Hub/Content/Sense_Hub/Visualizations/ScatterPl...
The threshold is not very practical as very often users need to analyze slightly more than that number of data points and they are stuck with the compressed view unless they make a selection or zoom in. Quite bothering is also the single color that can be applied to compressed data - so we have no real way to distinguish between different groups of points. With modern computing capabilities, it shouldn't be a problem to visualize 2k, 5k or 10k data points on a scatter plot.
Having a property to set the number of data points above which compression is applied will be the best approach. This can be implemented next to the current slider that sets compression resolution.