src/streamlit_app.py

import streamlit as st
import pandas as pd

# ---------------------------------------------------
# PAGE SETTINGS
# ---------------------------------------------------

st.set_page_config(
    page_title="Housing Price Estimator",
    page_icon="🏠",
    layout="centered"
)

# ---------------------------------------------------
# LOAD NEIGHBORHOOD SCORES
# ---------------------------------------------------

postal_scores = pd.read_csv("idf_postal_codes_neighborhood_scores.csv")

 postal_scores["postal_code"] = postal_scores["postal_code"].astype(str)


# ---------------------------------------------------
# MODEL COEFFICIENTS (from your R model)
# ---------------------------------------------------

INTERCEPT = -467418.10
COEF_SURFACE = 4932.35
COEF_ROOMS = -29954.53
COEF_NEIGHBORHOOD = 79383.79


def predict_price(surface, rooms, neighborhood_score):

    price = (
        INTERCEPT
        + COEF_SURFACE * surface
        + COEF_ROOMS * rooms
        + COEF_NEIGHBORHOOD * neighborhood_score
    )

    return price


# ---------------------------------------------------
# TITLE
# ---------------------------------------------------

st.title("🏠 Île-de-France Housing Price Estimator 2")

st.write(
"""
Estimate the price of a property using a **linear regression model**
trained on real estate data from Île-de-France.
"""
)

# ---------------------------------------------------
# USER INPUTS
# ---------------------------------------------------

st.sidebar.header("Property Information")

postal_code = st.sidebar.text_input(
    "Postal code",
    value="75001"
)

surface = st.sidebar.slider(
    "Surface (m²)",
    20,
    300,
    70
)

rooms = st.sidebar.slider(
    "Number of rooms",
    1,
    10,
    3
)

# ---------------------------------------------------
# FIND NEIGHBORHOOD SCORE
# ---------------------------------------------------

def get_neighborhood_score(postal_code):

    result = postal_scores[
        postal_scores["postal_code"] == postal_code
    ]

    if len(result) == 0:
        return None

    return result["neighborhood_score"].values[0]


# ---------------------------------------------------
# PREDICTION
# ---------------------------------------------------

if st.sidebar.button("Estimate price"):

    score = get_neighborhood_score(postal_code)

    if score is None:

        st.error("Postal code not found in the dataset.")

    else:

        price = predict_price(surface, rooms, score)

        st.subheader("Estimated Property Price")

        st.success(f"{int(price):,} €")

        st.write("Neighborhood score:", round(score, 2))


# ---------------------------------------------------
# MODEL FORMULA
# ---------------------------------------------------

st.write("---")

st.write("### Model Formula")

st.latex(
r"""
Price =
-467418
+ 4932 \times Surface
- 29954 \times Rooms
+ 79383 \times NeighborhoodScore
"""
)