Data Science Project Showcase

Notebook-based projects applying time series modeling, data analysis, and domain-informed forecasting using real-world datasets.

Each project folder is self-contained with reproducible environments and documented methodology.

Projects

• ice_cream_sales_analysis/ analysis.ipynb Exploratory deep-dive into sales patterns for an ice cream company, validating temporal assumptions and laying the groundwork for robust forecasting models.

• retail_sales_prophet_vs_sarimax/ forecasting.ipynb Comparative time series forecasting project evaluating Prophet and SARIMAX on multivariate retail sales data, with a focus on seasonal effects, holiday impacts, and short-term sales dynamics using exogenous variables.

• credit_card_fraud_detection/ main.ipynb Fraud detection showcase using PCA-transformed features and a Random Forest model. Includes business-aware metrics (fraud recovered, missed, and false alerts) and clear framing for executive stakeholders.

Tech Stack

• Core: Python 3.10+, Jupyter, pandas, NumPy
• Visualization: Seaborn, Matplotlib, Plotly
• Modeling: scikit-learn, statsmodels, Prophet
• Environment: pip + per project requirements.txt

Setup

We recommend installing dependencies inside a virtual environment (e.g., venv or conda).

Clone the repo and navigate into it:

git clone https://github.com/SoftwareSushi/data-science-notebooks.git

cd data-science-notebooks

Install project dependencies (each project has its own requirements.txt):

pip install -r <project-folder>/requirements.txt