Overview
What this work focused on
A weather-forecasting study using high-frequency campus sensor data to support short-term rainfall prediction.
Research / Time-Series Forecasting Research
Campus Weather Data Analysis for Short-Term Rainfall Forecasting
This project analyzed more than seven months of multivariate campus weather data collected at five-minute intervals, including temperature, humidity, wind, and rainfall signals. The work focused on preparing the dataset for time-series modeling through missing-data handling, feature transformation, and normalization, then supporting a CNN-based rainfall forecasting pipeline by organizing structured inputs for model training and evaluation.
Highlights
Key takeaways
Analyzed more than seven months of multivariate weather data collected at five-minute intervals, including temperature, humidity, wind, and rainfall.
Performed preprocessing work including missing-data handling, feature transformation, and normalization for time-series modeling.
Supported a CNN-based rainfall forecasting pipeline by preparing and analyzing structured input data for model training.