Data Scientist (Clean Energy)
CurrentUsing Machine Learning, Pandas, Python, R, SQL, and Statistical Inference to initiate projects such as those outlined below:Classification Models to discern Residential EV ownership from metering records. Built multi-class classification model (Random Forest and Logistic Regression) with seasonal AUC of 0.80 to 0.97 to detect residential high-rate electric vehicle charging from hourly metering records, thereby enabling the targeted marketing of Managed Charging and/or pre-emptive upgrading of relevant transformer infrastructure to avoid overloading and failure.Analysis of Potential Costs Savings from Managed Charging of Electric Vehicles. Built multiple regression model to predict wholesale power prices for the California and Mid-Continent electrical grids, allowing potential costs savings of 50% to be realized from EV charging, and Classification Model to predict the exact hours at which wholesale prices would be lowest.Analysis of the Shadow Carbon Footprint of Wind Energy. Built multiple regression model to quantify the increase in CO2 emissions, from aggregations at the individual plant level, due to the back-up of intermittent Wind Energy for the ERCOT electrical grid.
