🚲 Seoul Bike Sharing Demand Predictor

---

📌 Overview

This project focuses on predicting bike rental counts in Seoul using various regression models. The notebook explores both classical and neural network approaches to understand and forecast bike sharing demand based on weather and temporal features.

---

📊 Dataset

The Seoul Bike Sharing Demand Dataset is sourced from the UCI Machine Learning Repository:

Seoul Bike Sharing Demand [Dataset]. (2020). UCI Machine Learning Repository. https://doi.org/10.24432/C5F62R.

Key Features:

• Date, Hour, Temperature, Humidity, Wind Speed, Visibility, Dew Point, Solar Radiation, Rainfall, Snowfall, Bike Count, and more.

---

🔍 Notebook Contents

1️⃣ Dataset Information

• Description of the dataset and its variables.

2️⃣ Data Loading & Preprocessing

• Importing libraries (pandas, numpy, matplotlib, seaborn, scikit-learn, tensorflow).
• Loading and cleaning the dataset.
• Filtering for noon (hour 12) and dropping less relevant columns.

3️⃣ Exploratory Data Analysis (EDA)

• Scatter plots to visualize relationships between features and bike count.
• Dropping columns with weak linear relationships.

4️⃣ Data Splitting

• 60-20-20 split for training, validation, and test sets.

5️⃣ Modeling

• Linear Regression:
  • Simple Linear Regression (SLR) using only temperature.
  • Multiple Linear Regression (MLR) using all features.
• Neural Networks:
  • Linear regression with a single dense layer.
  • Multiple linear regression with neural nets (ReLU activations).
  • Complex neural networks for temperature-based prediction.

6️⃣ Visualization

• Training loss plots and model fit visualizations.

---

🛠️ How to Use

1. Open in Colab:

2. Run the Notebook:

   • Follow the step-by-step cells.
   • Ensure you have the required libraries installed.

3. Clone the Repo:

   git clone https://github.com/Divesh-Kshirsagar/Seoul-Bike-Sharing-Demand-Predictor.git

---

📜 License

This project is licensed under the MIT License – see the LICENSE file for details.

---