I think for my piezos those would come out as microvolts but I don't really know.
Well, I didn't spend a month on this project, but this what comes to my mind after reading:
1. Piezo is a current source. When you talking 20 mV or 1 V it doesn't make any sense w/o R-load. It's simple, sensor generate a current proportional to mechanical force, you setting resistor in parallel and getting voltage to amplify father.
2. There is a trade. Setting as high R as you could, you increasing voltage output ( V = I x R ). But the same time frequency response become sloppy, sensor would be high sensitive to single drops but completely lost in heavy drums, where pulses would "overlapped".
3. High R also source of noise.
4. You need to look at C (capacitance) in data sheet. Than, make estim how many water drops per second you need to count - max, it would gives you an Freq. Make a slack, 3-5 times coefficient. For example, 50 raindrops/sec = 50 Hz. x coeff 5 = 250 Hz.
5. You setting an R = F / C, where F from 4 and C from data sheet
6. Than you make a measurements of voltage, ( unfortunately, I'd not find Force-Current chart in data sheet), and based on that you make a search an amplifier that satisfy requirements : Gain, Noise, Freq., etc, (don't look at any crap some dumb landscape-designer post on-line),