This image [1] comes from "Converson to Islam in the Medieval Period" by Richard Bulliet [2]. The percentages inferred by Bulliet roughly track a sigmoidal curve. He makes an argument in analogy to the spread of a disease through a social network.

It should be noted that the deta is from Iran, but presumably the same arguments apply for Southeast Asia.

[1] http://4.bp.blogspot.com/-69BqAL9ARKg/T1WBF0NwxeI/AAAAAAAAAB...

[2] http://www.hup.harvard.edu/catalog.php?isbn=9780674732810

I doubt that there's enough data to plot. For early modern South East Asia, only a couple of records per century survive for the total populations of the main cities, and those are pretty dodgy.

There might not be any quantitative records of religious beliefs. When our ancestors encountered people of a different religion, it was more about chopping heads off than about counting them.