Project Highlights:

• Data Exploration: 🧹 Employed one-way ANOVA to identify features most relevant to rainfall prediction and utilized the IQR method to remove outliers, ensuring cleaner and more accurate data.
• Feature Engineering: 🛠️ Leveraged time series forecasting, and transformed date features into sine and cosine values for month and day to improve model performance. Utilized a column transformer to handle nearly 25 input columns effectively.
• Model Development: 🤖 Designed an artificial neural network (ANN) with a ReLU activation function, optimizing the model to deliver accurate rainfall predictions.

Challenges Faced:

• Handling Complex Data: 📊 With nearly 25 input columns, tracking and processing each feature was intricate. I employed a column transformer to manage this complexity and ensure smooth data input for the model.
• Deployment Difficulties: 🚀 Deploying the model posed its own set of challenges. I used Flask to deploy the model, which required careful handling of each input column and integration with the prediction system. The deployment process was a valuable learning experience, ensuring that the model could handle real-time inputs effectively.