Study and Implementation of Data Mining in Urban Gardening

Abstract

The system is essentially a three-part development, utilising Android, Java Servlets, and Arduino
platforms to create an optimised and automated urban-gardening system. Linking of these
platforms are through a five-step process – monitoring, recording, processing, optimising, and
reporting. The process begins through the monitoring of plants using sensors connected to the
Arduino device. Attached sensors generate data and send these data to the Java Servlet application
through a WIFI module. These data are processed and stored in appropriate formats in a MySQL
server database. Using the J48 tree algorithm implemented through WEKA API on a Java Servlet,
data provided is processed to derive a health index of the plant, with the possible outcomes set to
“Good,” “Okay”, or “Bad”. This information is then utilised to optimise the automated plantcaregiving features that the system contains, which are irrigation and sunlight through LED grow
lights. Feedback given to the user to inform them of methods by which they can improve their
plant’s health condition, derived through the information generated from the data-mining module.
A user can then remotely monitor and care for their plants. The major caregiving tasks of the
plants in this system is automated and its users are equipped with a powerful tool that informs and
educates them on the conditions of their plant, providing them with information that aids with
improvement of the plants’ health conditions.