🍗 KFC Siam — Customer Behavior Data Analysis

A multi-page Streamlit dashboard for analyzing customer behavior data collected from KFC Siam (Siam University). This project covers statistical summaries, data visualizations, and machine learning models built from survey responses.

📋 Survey form (data collection): https://survey.htunthihamyo.com/kfc/

---

👥 Team Members

Global Academy, Siam University — Year 2, Semester 2 · Data Science Project

---

📋 Overview

This dashboard analyses responses collected via the KFC Siam Customer Behavior Survey. It explores:

• Demographics — age, gender, nationality, occupation, faculty
• Spending behavior — budget ranges and patterns by group
• Ordering habits — visit frequency, dine type, order method, payment
• Menu & add-on preferences — most popular items
• Promotions — preferred promotion types
• Ratings — flavor quality and service satisfaction
• Machine learning models — budget and occupation prediction

---

🗂️ Project Structure

KFC final/
├── KFCfinal.py               # Entry point — Streamlit multi-page navigation
├── VisualizationforKFC.py    # Page 1: Key main charts & visualizations
├── Statistics.py             # Page 2: Statistical summaries (mode, median, etc.)
├── OtherFindings.py          # Page 3: Additional charts and correlation heatmap
├── DecisionTree.py           # Page 4: ML — Decision Tree (predict budget)
├── KNeighbour.py             # Page 5: ML — K-Nearest Neighbours (predict occupation)
├── KFC.xlsx                  # Survey dataset (Clean_data, Addon, Promotion sheets)
├── kfc_insights_export (1).xlsx  # Exported insights / summary data
└── README.md                 # This file

---

📊 Dashboard Pages

1 · Key Main Charts (VisualizationforKFC.py)

Interactive visualizations using Plotly and Matplotlib:

Chart	Description
Budget by Occupation	Bar chart showing spending distribution across student, staff, TA, others
Demographic Spending by Age	Budget breakdown per age group
Demographic Spending by Nationality	Stacked bar chart per nationality
Most Popular Main Menu	Horizontal histogram — Chicken dominates
Add-on Popularity	French fries are the top add-on
Order Type Distribution	Pie chart — Individual (40.5%) and Promotion (38.1%) lead
Order Method by Age Group	Kiosk preferred by younger customers

---

2 · Statistics (Statistics.py)

Descriptive statistics with interpretations:

• Gender & Nationality distribution with frequency tables
• Dine type & Payment mode analysis
• Age — Median & mode: 18–22 age group
• Budget — Median & mode: 100–199 THB range
• Visit Frequency — Median & mode: Sometimes
• Ratings — Flavor and service rating medians & modes

---

3 · Other Findings (OtherFindings.py)

Additional exploratory charts:

Chart	Key Finding
Budget Distribution	Most customers spend 100–199 THB
Budget vs Occupation Boxplot	TAs spend the most overall
Ordering Method Histogram	Kiosk is the most preferred ordering method
Budget by Age Boxplot	28–35 group shifts toward 300+ budget tier
Promotion Preference	Discounts are the #1 preferred promotion
Order Type by Age	18–22 group dominates individual and promotion orders
Major Distribution	Global Academy students = 74.5% of respondents
Correlation Heatmap	Age, budget, visit frequency are strongly correlated (0.92–0.98)

---

4 · Decision Tree (DecisionTree.py)

Goal: Predict a customer's spending budget category.

• Features: Order type, age, occupation
• Algorithm: DecisionTreeClassifier (max depth = 3)
• Preprocessing: StandardScaler + LabelEncoder
• Interactive: Adjustable test/train ratio; real-time prediction on user input

---

5 · K-Nearest Neighbours (KNeighbour.py)

Goal: Predict a customer's occupation category.

• Features: Age, budget, order type, order method
• Algorithm: KNeighborsClassifier
• Interactive: Selectable K value (1–15), adjustable test ratio, real-time prediction

---

🛠️ Tech Stack

Technology	Purpose
Python 3	Core language
Streamlit	Multi-page web dashboard
Pandas	Data loading & manipulation
NumPy	Numerical operations
Matplotlib & Seaborn	Static charts
Plotly Express	Interactive charts
scikit-learn	Machine learning models

---

🚀 Getting Started

Prerequisites

Install the required Python packages:

pip install streamlit pandas numpy matplotlib seaborn plotly scikit-learn openpyxl

Run the dashboard

streamlit run KFCfinal.py

The app will open at http://localhost:8501 in your browser.

Note: KFC.xlsx must be in the same directory as the scripts. The workbook must contain three sheets: Clean_data, Addon, and Promotion.

---

📁 Dataset (KFC.xlsx)

Sheet	Contents
Clean_data	Main cleaned survey responses (demographics, behavior, ratings, encoded columns)
Addon	Add-on item counts
Promotion	Promotion type counts

Key Columns (Clean_data)

Column	Description
age	Age range label
age_enc	Encoded age (1–5)
gender	Male / Female / LGBTQ+
nationality	Country of origin
occupation	Student / Staff / TA / Others
major	Faculty (students only)
budget	Spending budget label
budget_enc	Encoded budget (1–4)
visitFrequency_enc	Encoded visit frequency (1–5)
dineType	Dine-in / Takeaway / Delivery
orderType	Individual / Promotion / Group / Snack Sharing
orderMethod	Counter / Kiosk / App
payment	Cash / PromptPay / Card / True Money
menuCategory	Most purchased menu category
flavorRating	Food flavor rating (1–5)
serviceRating	Service experience rating (1–5)

---

📌 Key Insights

• 🏆 Chicken is overwhelmingly the most popular menu category
• 🍟 French fries are the most preferred add-on
• 💸 Most customers spend 100–199 THB per visit
• 👥 The majority of respondents are 18–22 year old students
• 🖥️ Kiosk is the most popular ordering method
• 🏷️ Discounts are the most preferred promotion type
• 📈 Strong correlation (0.92–0.98) exists between age, budget, and visit frequency

---

© 2026 KFC Siam · Data Science Research Project · Global Academy, Siam University