How Real Estate Companies Use Data Science for Business Growth

import pandas as pd
from sklearn.model_selection import train_test_split
from sklearn.compose import ColumnTransformer
from sklearn.preprocessing import OneHotEncoder
from sklearn.pipeline import Pipeline
from sklearn.ensemble import GradientBoostingRegressor
from sklearn.metrics import mean_absolute_error

df = pd.read_csv("listings.csv")
X = df[["carpet_sqft", "bedrooms", "age_years", "locality", "parking"]]
y = df["price"]

prep = ColumnTransformer([("loc", OneHotEncoder(handle_unknown="ignore"), ["locality"])],
                         remainder="passthrough")
model = Pipeline([("prep", prep), ("gbr", GradientBoostingRegressor(random_state=42))])

X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=0.2, random_state=42)
model.fit(X_tr, y_tr)
print("MAE:", mean_absolute_error(y_te, model.predict(X_te)))

df["price_per_sqft"] = df["price"] / df["carpet_sqft"]
print(df.groupby("locality")["price_per_sqft"]
        .median()
        .sort_values(ascending=False)
        .head(10))

from sklearn.linear_model import LogisticRegression
from sklearn.model_selection import train_test_split
from sklearn.metrics import roc_auc_score

leads = pd.read_csv("leads.csv")
X = pd.get_dummies(leads[["source", "budget_band", "site_visit", "hours_to_contact"]],
                   columns=["source", "budget_band"], drop_first=True)
y = leads["converted"]
X_tr, X_te, y_tr, y_te = train_test_split(X, y, test_size=0.25, random_state=1, stratify=y)
clf = LogisticRegression(max_iter=1000).fit(X_tr, y_tr)
print("AUC:", roc_auc_score(y_te, clf.predict_proba(X_te)[:, 1]))
leads["score"] = clf.predict_proba(X)[:, 1]

leads["enquiry_date"] = pd.to_datetime(leads["enquiry_date"])
monthly = leads.set_index("enquiry_date").resample("MS").size()
rolling = monthly.rolling(3).mean()
print(pd.DataFrame({"enquiries": monthly, "3m_avg": rolling}).tail(12))

import numpy as np

def haversine_km(lat1, lon1, lat2, lon2):
    r = 6371.0
    p1, p2 = np.radians(lat1), np.radians(lat2)
    dphi = p2 - p1
    dlmb = np.radians(lon2 - lon1)
    a = np.sin(dphi / 2) ** 2 + np.cos(p1) * np.cos(p2) * np.sin(dlmb / 2) ** 2
    return 2 * r * np.arcsin(np.sqrt(a))

# distance from each listing to one station
df["km_to_station"] = haversine_km(df["lat"], df["lon"], 28.5850, 77.3110)

Area scorecard (score 1-5 each)
Area: ____________
Connectivity (metro / highway)   __
Schools and hospitals nearby     __
Price trend (last 3 years)       __
Upcoming supply (lower = better) __
Employment hubs nearby           __
Total                            __ / 25

Project: Property price estimator for one city
Steps:
1. Collect a public listings dataset (or synthetic data)
2. Clean: remove duplicates, fix units, handle outliers
3. Explore: price per sq ft by locality, area vs price
4. Features: locality, area, bedrooms, age, amenities
5. Model: baseline (median by locality) vs gradient boosting
6. Evaluate with MAE on a hold-out set
7. Explain top features in plain English

# A simple baseline every model must beat
train = X_tr.assign(price=y_tr)
median_ppsf = (train["price"] / train["carpet_sqft"]).groupby(train["locality"]).median()
overall = (train["price"] / train["carpet_sqft"]).median()
pred = X_te["locality"].map(median_ppsf).fillna(overall) * X_te["carpet_sqft"]
print("Baseline MAE:", mean_absolute_error(y_te, pred))

Findings (fill with your results):
- Model MAE vs baseline MAE: ____ vs ____
- Top 3 price drivers: ____, ____, ____
- Localities where the model is least accurate: ____
- Next step: add distance to metro and amenities