☕ Coffee Sales Analysis & Prediction

An interactive data analysis and prediction project using coffee shop transaction data (March 2024 → March 2025).
The project includes EDA, statistical analysis, regression models, and a Streamlit dashboard.

---

📂 Project Structure

├── Coffe_sales.csv                        # Dataset
├── Project_coffe_sales_with_models.ipynb  # Jupyter Notebook (EDA + Modeling)
├── dashboard.py                                 # Streamlit app for dashboard
├── requirements.txt                       # Dependencies
└── README.md                              # Project documentation

---

🎯 Objectives

• Analyze sales patterns by time, weekday, month, and season.
• Identify top-selling coffee types and customer spending habits.
• Compare weekday vs weekend performance.
• Build predictive models (OLS, Linear Regression, XGBoost).
• Deploy results via a Streamlit dashboard.

---

📊 Dataset

• Source: Kaggle Coffee Shop Transactions Dataset
• Period Covered: March 2024 → March 2025 (1 full year)
• Transactions: 3,500

Column	Description
hour_of_day	Hour of purchase (0–23)
cash_type	Payment type (Cash/Card)
money	Transaction amount
coffee_name	Coffee type (Latte, Americano, etc.)
Time_of_Day	Morning / Afternoon / Night
Weekday	Day of the week
Month_name	Month name
Weekdaysort	Numeric weekday (1=Mon … 7=Sun)
Monthsort	Numeric month (1=Jan … 12=Dec)
Date	Transaction date
Time	Transaction time

📌 Dataset Link

---

🛠️ Installation

Clone the repository:

git clone https://github.com/MUSAB10000/Project-Coffee-Sales.git
cd Project-Coffee-Sales

---

📈 Key Insights

Sales are evenly split between weekdays and weekends (~50/50).

Autumn recorded the highest seasonal sales (~29%).

Americano with Milk and Latte are the most popular coffee types.

XGBoost provided the best predictive accuracy compared to OLS and Linear Regression.