Predicting House Prices Using Decision Trees in Python (Beginner to Pro Guide)

Last Updated on May 3, 2026 by Rajeev Bagra

Learn how to predict house prices using Decision Tree Regressor in Python with step-by-step code, real-world insights, and practical business applications.

ߓ Introduction

Predicting house prices is one of the most practical applications of machine learning in real estate, finance, and business strategy.

In this guide, we’ll use the Melbourne Housing Dataset and build a predictive model using a DecisionTreeRegressor to estimate property prices based on features like rooms, location, and land size.

.

ߏ️ Understanding the Dataset

The dataset includes:

ߛ️ Rooms
ߛ Bathrooms
ߓ Landsize
ߏ️ Building Area
ߓ Year Built
ߌ Latitude & Longitude
ߒ Price (Target Variable)

⚙️ Step 1: Load and Prepare Data

import pandas as pd

melbourne_file_path = '../input/melbourne-housing-snapshot/melb_data.csv'
melbourne_data = pd.read_csv(melbourne_file_path)

filtered_melbourne_data = melbourne_data.dropna(axis=0)

y = filtered_melbourne_data.Price

features = ['Rooms', 'Bathroom', 'Landsize', 'BuildingArea', 
            'YearBuilt', 'Lattitude', 'Longtitude']

X = filtered_melbourne_data[features]

Step 2: Build the Model

from sklearn.tree import DecisionTreeRegressor

melbourne_model = DecisionTreeRegressor()
melbourne_model.fit(X, y)

Step 3: Make Predictions

print(melbourne_model.predict(X.head()))

This gives predicted prices for sample houses.

Step 4: Validate Your Model

from sklearn.model_selection import train_test_split
from sklearn.metrics import mean_absolute_error

train_X, val_X, train_y, val_y = train_test_split(X, y, random_state=0)

melbourne_model.fit(train_X, train_y)
predictions = melbourne_model.predict(val_X)

mae = mean_absolute_error(val_y, predictions)
print(mae)

ߑ Mean Absolute Error (MAE) tells you how wrong your predictions are on average.

How Decision Trees Work

A DecisionTreeRegressor works like a flowchart:

Splits data based on feature values
Creates decision rules
Predicts outcomes at leaf nodes

Example:

If Rooms > 3 → higher price
If Landsize < 200 → lower price

⚠️ Common Pitfall: Overfitting

A fully grown tree memorizes data instead of learning patterns.

Fix it using:

melbourne_model = DecisionTreeRegressor(max_depth=5)

ߒ Business & Economic Relevance

This model is not just academic—it has real-world impact:

ߏ Real Estate Platforms
Estimate property values instantly
Power recommendation engines
ߒ Investors
Identify undervalued properties
Optimize buying decisions

ߏ Banks & Lenders
Assess collateral value
Reduce loan risk
ߓ Economic Insight
Understand urban development trends
Track pricing patterns across regions

ߚ Going Beyond: Better Models

Decision Trees are great for learning, but for production:

ߌ Random Forest → More accurate, less overfitting
⚡ Gradient Boosting → State-of-the-art performance

ߧ Final Thoughts

You’ve just built a complete machine learning pipeline:

✅ Data loading
✅ Feature selection
✅ Model training
✅ Prediction
✅ Evaluation

This is the foundation of applied AI in business.

---

Reference: https://www.kaggle.com/learn/intro-to-machine-learning