Design and Implementation of Weather Monitoring and Prediction System

For the last few years, a lot of research works are carried out on different intelligent technologies like Internet of Things (IoT) and machine learning. Their implementation is possible through a variety of IoT hardware platforms like ESP8266 chip. In the paper we have proposed a method for automatically weather monitoring system that has access to the dynamic as well as real-time atmospheric data within a specific range. IoT, embedded systems, along with the machine learning concepts provide foundation of this weather monitor system. Additionally, the system uses wireless technologies like sensors and other electrical gadgets. Primary objective of the proposed system is to use sensors for sensing different atmospheric conditions like temperature, humidity, rainfall and air pressure. Data has been collected through the sensors and uploaded to ThingSpeak server. This information is also displaced on a webpage from where users may view them. Machine learning environment is built up using different methods like the Logistic Regression model, KNN, SVM, Gaussian naivebayes, XGBoost and ANN. These models are trained based on previous records that are available from different sensors. Additionally, NodeMCU collects sensor data, including temperature, humidity, rainfall and pressure and is uploaded to a ThingSpeak cloud server. These data are then stored in a comma separated Value (CSV) file. Later it is transferred to a Jupyter notebook which runs on Python environment. In this experiment only, real-time data collected from different sensors has been used.