Somebody could use this as a starting point. http://touchscale.co/ You'd have to collect new data on touch strength vs. weight to get the regression parameters.
(If you do this, let me know and I can add it to the site above, and then we can both delight in the surprisingly large amount of unmonetizable traffic it gets.)
(If you do this, let me know and I can add it to the site above, and then we can both delight in the surprisingly large amount of unmonetizable traffic it gets.)