Abstract:
This paper introduces a real-time water quality dataset of
five ponds for fish farming obtained through an IoT frame-
work for monitoring the aquatic environmental conditions.
It utilizes sensors and an Arduino microcontroller to col-
lect data on pH, temperature, and turbidity in pond water
in Jamalpur District, Bangladesh. The data is stored in an
IoT cloud platform named ThingSpeak and analyzed using
10 machine learning algorithms. The dataset consists of 4
columns and 40,280 rows, where pH, temperature, turbid-
ity, and fish are recorded. Fish represents the target variable,
while the others serve as independent variables. Within the
dataset, there are 11 distinct fish categories including sing,
silver carp, Katla, prawn, karpio, shrimp, rui, pangas, tilapia,
magur, and koi. Results showed that only three ponds are
suitable for fish farming among five ponds and the Random
Forest algorithm performs the best. The study also includes
details of the IoT system’s hardware. This dataset will be use-
ful for researchers and fish farmers to predict fish survival.
