In 2012 a friend of mine approached me about building devices to harvest wind energy from freeways. My first thoughts went to assessing how much wind cars actually produce, so I built an anemometer which also contained a battery operated data-logger so that I could collect data and analyze the feasibility of the endeavor. This is really similar to what I did at Intel: collect reams of data and dive into it looking for trends or metrics. The results were pretty nifty: we collected weeks of data from freeways in Connecticut which showed very distinct trends over time, day to night. THe fun part was building a rather compleicated device from parts at Home Depot, and successfully ddeploying and retrieving it from a public place exposed to weather and traffic. I learned a lot about field reliability, especially when putting the device in the hands of non-techies.