The falling costs of battery storage and photovoltaic systems have substantially increased the number of "solar-plus-battery" installations in homes and buildings. The solar-plus-battery system enables homeowners to protect their homes during a power outage and save on their electricity bills by stacking multiple value streams that battery storage can provide. In this paper, we present EnergyBoost, a system that proactively controls battery charge and discharge operations, and investigate whether it makes sense economically to install a battery controlled by this system in different jurisdictions with distinct tariff structures. EnergyBoost solves an optimal control problem over a finite time horizon relying on physical models of a solar inverter and a lithium-ion battery, and supervised learning models for predicting the next day available solar energy and household demand. We propose two learning-based control algorithms for EnergyBoost, namely model predictive control and advantage actor-critic. We implement these algorithms on a Raspberry Pi and compare their performance with a rule-based controller under various pricing schemes using real traces of solar irradiance and power consumption of 70 homes located in the same jurisdiction. Our results indicate that EnergyBoost'S control policy outperforms the baseline policies in terms of reducing the average monthly electricity bill, yielding a bill that is, on average, only 7.6% worse than the best bill that can be theoretically achieved.