Public toilets experience the perpetual problem of lack of hygiene. Appointing an all-Time janitor is not possible at each and every toilet. Automation can reduce the janitor's burden by looking after the maintenance of the toilet and sending timely updates. We've tried to tackle this issue by developing an IoT-based smart system. We've used sensors to measure water level in tanks, water usage and to detect the presence of a person in the toilet. With the data from sensors, our system predicts the cleanliness of toilets. The sensors are interfaced with Raspberry Pi, which processes the sensor data and uploads it to the cloud. The necessary action is initiated by the Pi. The data collected by the sensors is fed to ThingSpeak, which generates graphs. With its ready-made machine learning tools, ThingSpeak identifies patterns in data and provides an analysis, which can uncover some useful information. © 2019 IEEE.