🌦️ London Weather Predictor

This project predicts London’s daily temperature using machine learning.
It follows a structured ML workflow with EDA, data cleaning, model training, and experiment tracking with MLflow.

---

📂 Project Structure

london-weather-predictor/
├── data/
│   ├── raw/                # Original datasets
│   │   └── london\_weather.csv
│   ├── clean/              # Processed/clean datasets
│   │   └── london\_weather\_clean.csv
│   └── mlruns/             # MLflow experiment logs
├── notebooks/
│   ├── eda.ipynb           # Data exploration and cleaning
│   └── train.ipynb         # Model training & evaluation
├── src/
│   └── data/
│       └── clean.py        # Data cleaning script
└── README.md

---

⚙️ Setup

1. Clone this repo:

   git clone https://github.com/joseph-ajegetina/london-weather-predictor.git
   cd london-weather-predictor```
   

2. Create a virtual environment and install dependencies:

   python -m venv venv
   source venv/bin/activate   # On Linux/Mac
   venv\Scripts\activate      # On Windows
   
   pip install -r requirements.txt

3. Make sure the dataset is in data/raw/london_weather.csv.

---

🧹 Data Cleaning

Run the cleaning script to generate the processed dataset:

---

📊 Notebooks

• EDA: explore dataset, visualize trends, and verify missing values

  jupyter notebook notebooks/eda.ipynb

• Training: train and evaluate models with MLflow tracking

  jupyter notebook notebooks/train.ipynb

---

🧪 Experiment Tracking (MLflow)

MLflow logs are stored in:

data/mlruns/

Start MLflow UI

mlflow ui --backend-store-uri file:data/mlruns

Open http://127.0.0.1:5000 to browse experiments.

---

🚀 Workflow

1. EDA (eda.ipynb)

   • Explore dataset (trends, correlations, missing values)
   • Feature selection & preprocessing

2. Training (train.ipynb)

   • Train multiple models:
     • Linear Regression
     • Decision Tree Regressor
     • Random Forest Regressor
   • Evaluate using RMSE, MAE, R²
   • Log experiments with MLflow

---

📊 Example Outputs

Correlation Heatmap

Shows relationships between weather variables and mean temperature.

Temperature Trend

Monthly average temperatures across years.

ML Comparison

Metric value for the various models

---

🔧 Tech Stack

• Python (pandas, numpy, matplotlib, seaborn, scikit-learn)
• MLflow for experiment tracking
• Jupyter Notebooks for development

---

📈 Next Steps

• Add hyperparameter tuning (Grid/Random Search)
• Try Gradient Boosting / XGBoost / LightGBM
• Deploy best model as an API (FastAPI/Flask)