New to this whole blog thing so please excuse any spelling / grammar issues. 
My smarter life project idea is an intelligent house plant watering system. Initial goal is to build a system that can monitor soil moisture of my small collection of house plants and dispense water as needed. Currently that typically only gets done when I think of it which is usually after there is major wilting. The system will need to be self sufficient or at the very least be able to alert me when maintenance (filling the water container for example) needs to be done.
Other secondary goals include ambient light tracking (both for charting and for disabling alarms and watering at night), temperature tracking, and possibly automation of the window shades.
So far on the project front I have done some research into moisture probes and there are a few different designs available.
Resistive:
Typically a simple set of metal rods installed in the pot and the resistance of the soil is measured. Downsides to this type are the rods will corrode over time and need to be cleaned / replaced. (if I can't remember to water the plants now not going to remember to clean electrodes) Also the readings are highly sensitive to the composition and temperature of the soil.
Galvanic:
Two dissimilar metal rods are installed in the soil and a voltage level is produced between them based on the amount of electrolyte (soil + water) is present. Again have all the same drawbacks as resistive sensors.
Capacititive:
This is the first of a series of "non contact" sensors. Two plates are isolated from each other and the soil by a insulator. Depending on the amount of water present in the soil the dielectric changes which results in a change of the capacitance value. Some downsides are the need to calibrate to the soil conditions. So far this is my preferred method due to the low maintenance requirements and only medium complexity.
Others:
There are quite a few other systems out there including TDR, TDT, volumetric, and even a thermal method (similar to how a mass flow meter in a car engine works). For now I am going to concentrate on the capacitive meter due to the higher complexity introduced with these methods.
Other than some research pickup up a bag of potting soil for making up some sensor test samples. Plan it to have one that is moderate, Dry one batch out in the oven for the max dry sample, and heavily water the last sample. This way I can test the performance of the various sensor configurations / types easily on my bench without abusing my plants too much.