This project analyses a dataset of restaurants in Sydney (2018).
It applies the full data science workflow: Exploratory Data Analysis (EDA), Predictive Modelling (Regression & Classification), Geospatial Visualisation, and Reproducibility with Git, Git LFS, and DVC.
- Clone the repository
git clone https://github.com/MuhammadAhmad-Flutter-Developer/DataScienceTechnologyAndSystem_Assignmnet1/tree/main cd assignment1-eatingout - Create and activate a virtual environment (recommended) python -m venv .venv
.venv\Scripts\activate
source .venv/bin/activate
3. Install dependencies
pip install -r requirements.txt
- Run Jupyter Notebook
The full analysis is inside the notebook:
jupyter notebook git.ipynb
- Reproduce Pipeline with DVC
If using DVC stages, you can reproduce the workflow with:
dvc repro
- Run PySpark Models
PySpark regression and classification models are included in the notebook. Make sure Java and Spark are installed before running those cells.
📊 Results to Expect
EDA
Missing values handled and summary statistics reported.
Distribution of cost, ratings, and restaurant types.
Top suburbs with the most restaurants.
Geospatial maps showing cuisine density across Sydney.
Interactive Plotly visualisation comparing costs of "Excellent" vs "Poor" rated restaurants.
Regression Models
Linear Regression and Gradient Descent compared using Mean Squared Error (MSE).
PySpark Linear Regression benchmarked for scalability.
Classification Models
Binary classification of ratings (Poor/Average vs Good+).
Logistic Regression, Random Forest, Gradient Boosted Trees, and SVM compared (F1, Precision, Recall).
PySpark Logistic Regression evaluated with class balancing.
Reproducibility
Git used for version control.
Git LFS tracks large files (datasets, visualisations, models).
DVC manages dataset and modelling pipeline for reproducibility.
👤 Author
Muhammad Ahmad Master of Data Science, University of Canberra