Let’s turn our real estate price prediction model into a working web application.
Option 1: Flask Web App (Python + HTML)
A lightweight setup ideal if you want full control over frontend design.
What You will Build:
A form to enter house details
A backend that loads your trained model and
returns the predicted price
Step 1: Save the Trained Model
In your existing Python script, after training:
import joblib
# Save the trained model
joblib.dump(model, 'price_model.pkl')
Step 2: Project Structure
real_estate_app/
│
├── app.py # Flask backend
├── price_model.pkl # Trained model
├── templates/
│ └── index.html # Frontend form
└── static/
└── style.css # Optional styling
Step 3:
app.py
(Flask Backend)
from flask import Flask, render_template, request
import pandas as pd
import joblib
app = Flask(__name__)
model = joblib.load('price_model.pkl')
# Update this list with all location dummy variables used in training
location_columns = ['Location_New York', 'Location_San Francisco', 'Location_Chicago']
@app.route('/')
def home():
return render_template('index.html')
@app.route('/predict', methods=['POST'])
def predict():
data = request.form
size = float(data['size'])
bedrooms = int(data['bedrooms'])
bathrooms = int(data['bathrooms'])
year_built = int(data['year_built'])
location = data['location']
# Initialize input with zeroes for all locations
input_data = {
'Size (sqft)': size,
'Bedrooms': bedrooms,
'Bathrooms': bathrooms,
'Year Built': year_built,
}
for loc in location_columns:
input_data[loc] = 1 if loc == f'Location_{location}' else 0
input_df = pd.DataFrame([input_data])
prediction = model.predict(input_df)[0]
return render_template('index.html', prediction=f"${prediction:,.2f}")
if __name__ == '__main__':
app.run(debug=True)
templates/index.html
(Form + Result)
<!DOCTYPE html>
<html>
<head>
<title>Real Estate Price Predictor</title>
</head>
<body>
<h2>Enter Property Details</h2>
<form method="POST" action="/predict">
<label>Size (sqft):</label><input name="size" required><br>
<label>Bedrooms:</label><input name="bedrooms" required><br>
<label>Bathrooms:</label><input name="bathrooms" required><br>
<label>Year Built:</label><input name="year_built" required><br>
<label>Location:</label>
<select name="location">
<option value="New York">New York</option>
<option value="San Francisco">San Francisco</option>
<option value="Chicago">Chicago</option>
</select><br>
<button type="submit">Predict Price</button>
</form>
{% if prediction %}
<h3>Predicted Price: {{ prediction }}</h3>
{% endif %}
</body>
</html>
Option 2: Streamlit Web App
Fastest way to build interactive ML dashboards – no HTML required.
app.py
(Streamlit Script)
import streamlit as st
import pandas as pd
import joblib
# Load the model
model = joblib.load('price_model.pkl')
# Define all possible locations used in training
location_columns = ['Location_New York', 'Location_San Francisco', 'Location_Chicago']
st.title(" Real Estate Price Predictor")
size = st.number_input("Size (sqft)", value=1000)
bedrooms = st.number_input("Bedrooms", min_value=0, step=1, value=3)
bathrooms = st.number_input("Bathrooms", min_value=0, step=1, value=2)
year_built = st.number_input("Year Built", min_value=1800, max_value=2025, value=2000)
location = st.selectbox("Location", ["New York", "San Francisco", "Chicago"])
# Prepare input data
input_data = {
'Size (sqft)': size,
'Bedrooms': bedrooms,
'Bathrooms': bathrooms,
'Year Built': year_built,
}
for loc in location_columns:
input_data[loc] = 1 if loc == f'Location_{location}' else 0
input_df = pd.DataFrame([input_data])
if st.button("Predict Price"):
prediction = model.predict(input_df)[0]
st.success(f"Predicted Price: ${prediction:,.2f}")
To Run the Streamlit App:
streamlit run app.py
Requirements File (requirements.txt)
For either setup, include:
pandas
scikit-learn
joblib
flask # For Flask version
streamlit # For Streamlit version
Recommendation
Use Streamlit if you want fast development, live interaction, and easy UI
Use Flask if you want more control over frontend and possibly host it on a website with more custom UX