An Occupancy Sensor is an indoor device which detects a person’s presence in a closed environment. Environmental sensors, which possess capabilities to sense the quality of air indoors, are affixed in a room in a big mall. These sensors could range from basics such as air quality and humidity sensor to a complete measurement system; which, with its array of various sensors, measures the air quality indoors. The room may also be affixed with an array of networked sensors. Such sensors primarily measure the indoor-air quality and the capabilities of these sensors in detecting are occupancy largely overlooked. Detecting occupancy involves techniques which ascertain the presence of people, quantify the number of occupants, and also determine any non-living objects. There are numerous environmental sensors that are employed in detection of various gases such as carbon dioxide (CO2) and Total volatile organic compounds (TVOC) sensors. Such ones can be affixed for detection of gases inside a shopping mall. The indoor-air quality can be considerably enhanced once we are able to detect and measure the gases in a room. CO2 standalone sensor detects only the composition of carbon dioxide. On the other hand, TVOC is already contains a CO2 sensor and can also sense other gases. Classification of occupancy detection can be done with a variety of machine learning algorithms. Earlier research used naive Bayes classifier to detect occupancy with a tool named Weka. In this paper, we explore the implementation of Random-forest, a machine learning algorithm, to detect occupants.