🏪 Location Decision Tree — Will the Business Succeed?

🔍 Overview

A classroom Decision Tree activity that predicts whether a small business location will be successful based on three intuitive features:

Feature	Meaning
Person	Foot traffic — number of people passing per hour
Income	Average income in the area (in thousands)
Shops	Number of competing shops nearby

The label is Successful → Yes / No.

---

📦 Features

• Built-in sample dataset (15 rows) — no external data needed to start
• Auto-creates location.csv on first run
• Trains a DecisionTreeClassifier with a configurable max_depth
• Reports test-set accuracy
• NEW: All key parameters exposed as command-line options
• NEW: Optional tree visualization saved as a PNG (--save-plot)
• NEW: --regenerate flag to refresh the CSV from the built-in sample
• 🐛 Bug fix: previous version used nested-list column headers ([["Person",...]]) which broke pandas indexing — fixed to a flat list

---

🧠 Dataset (built-in sample)

Person	Income	Shops	Successful
20	22	3	Yes
30	17	8	No
10	30	2	Yes
50	15	3	Yes
60	16	5	Yes
5	20	4	No
...	...	...	...

15 rows total. Saved to location.csv automatically the first time you run the script.

---

🧰 Requirements

pip install -r requirements.txt

---

⚙️ How to Run

Default run (creates location.csv if missing, trains, and prints accuracy):

python Location_DT_Activity.py

Custom run:

python Location_DT_Activity.py --csv location.csv --max-depth 3 --test-size 0.25 --seed 42 --save-plot tree.png

Regenerate the CSV from the built-in sample:

python Location_DT_Activity.py --regenerate

---

🛠️ Command-Line Options

Flag	Default	Description
--csv	location.csv	Path to the CSV dataset (auto-created if missing)
--max-depth	2	Max depth of the decision tree
--test-size	0.25	Fraction of data used for testing (0.0–1.0)
--seed	None	Random seed for reproducible splits
--save-plot	off	If set, saves a tree visualization PNG
--regenerate	off	Overwrite the CSV with the built-in sample before training

---

📊 Example Output

Created sample dataset at: location.csv

The dataset:
    Person  Income  Shops Successful
0       20      22      3        Yes
1       30      17      8         No
...

The predictions of the model are:
 ['Yes' 'No' 'Yes' 'Yes']

Accuracy = 0.7500

---

🎓 Learning Objectives

1. Build a DataFrame from a Python list of records
2. Save / load CSV files with pandas
3. Separate features from the label
4. Use train_test_split for a fair evaluation
5. Train a DecisionTreeClassifier and reason about max_depth
6. Interpret the resulting tree visually (--save-plot)

---

📁 Project Structure

Location-Decision-Tree-/
├── Location_DT_Activity.py   # Main script
├── location.csv              # Auto-generated on first run
├── requirements.txt          # Python dependencies
└── README.md

---

📄 License

Educational use — Credits to CS42.org.