Data Science for Real Estate: Housing Price Prediction
OUYANG ZIXUAN (24215697)
03 January 2026
1 Introduction
The real estate industry frequently struggles with inaccurate
property valuation, driven by rapidly changing market
conditions and heterogeneous property characteristics.
This report demonstrates how Data Science and Machine
Learning can be applied to:
- model the relationship between housing characteristics and price,
and
- build a predictive model to support data-driven valuation decisions.
We use a real estate dataset containing structural attributes (area, number of bedrooms, bathrooms, stories) and qualitative features (road access, furnishing status, etc.) to build and evaluate predictive models.
2 Load Libraries and Dataset
library(tidyverse)
library(caret)
library(ggplot2)
library(randomForest)
library(Metrics)
library(corrplot)
# Load dataset
df <- read.csv("/Users/oyzx/Downloads/Housing_Price_Data.csv")
# Preview column names
names(df)## [1] "price" "area" "bedrooms" "bathrooms"
## [5] "stories" "mainroad" "guestroom" "basement"
## [9] "hotwaterheating" "airconditioning" "parking" "prefarea"
## [13] "furnishingstatus"## price area bedrooms bathrooms stories mainroad guestroom basement
## 1 13300000 7420 4 2 3 yes no no
## 2 12250000 8960 4 4 4 yes no no
## 3 12250000 9960 3 2 2 yes no yes
## 4 12215000 7500 4 2 2 yes no yes
## 5 11410000 7420 4 1 2 yes yes yes
## 6 10850000 7500 3 3 1 yes no yes
## hotwaterheating airconditioning parking prefarea furnishingstatus
## 1 no yes 2 yes furnished
## 2 no yes 3 no furnished
## 3 no no 2 yes semi-furnished
## 4 no yes 3 yes furnished
## 5 no yes 2 no furnished
## 6 no yes 2 yes semi-furnishedWe observe key variables such as:
price: selling price of the house
area: built-up area
bedrooms,bathrooms,stories
mainroad,guestroom,basement,hotwaterheating,airconditioning,parking,prefarea,furnishingstatus
3 Data Understanding
## 'data.frame': 545 obs. of 13 variables:
## $ price : int 13300000 12250000 12250000 12215000 11410000 10850000 10150000 10150000 9870000 9800000 ...
## $ area : int 7420 8960 9960 7500 7420 7500 8580 16200 8100 5750 ...
## $ bedrooms : int 4 4 3 4 4 3 4 5 4 3 ...
## $ bathrooms : int 2 4 2 2 1 3 3 3 1 2 ...
## $ stories : int 3 4 2 2 2 1 4 2 2 4 ...
## $ mainroad : chr "yes" "yes" "yes" "yes" ...
## $ guestroom : chr "no" "no" "no" "no" ...
## $ basement : chr "no" "no" "yes" "yes" ...
## $ hotwaterheating : chr "no" "no" "no" "no" ...
## $ airconditioning : chr "yes" "yes" "no" "yes" ...
## $ parking : int 2 3 2 3 2 2 2 0 2 1 ...
## $ prefarea : chr "yes" "no" "yes" "yes" ...
## $ furnishingstatus: chr "furnished" "furnished" "semi-furnished" "furnished" ...## price area bedrooms bathrooms
## Min. : 1750000 Min. : 1650 Min. :1.000 Min. :1.000
## 1st Qu.: 3430000 1st Qu.: 3600 1st Qu.:2.000 1st Qu.:1.000
## Median : 4340000 Median : 4600 Median :3.000 Median :1.000
## Mean : 4766729 Mean : 5151 Mean :2.965 Mean :1.286
## 3rd Qu.: 5740000 3rd Qu.: 6360 3rd Qu.:3.000 3rd Qu.:2.000
## Max. :13300000 Max. :16200 Max. :6.000 Max. :4.000
## stories mainroad guestroom basement
## Min. :1.000 Length:545 Length:545 Length:545
## 1st Qu.:1.000 Class :character Class :character Class :character
## Median :2.000 Mode :character Mode :character Mode :character
## Mean :1.806
## 3rd Qu.:2.000
## Max. :4.000
## hotwaterheating airconditioning parking prefarea
## Length:545 Length:545 Min. :0.0000 Length:545
## Class :character Class :character 1st Qu.:0.0000 Class :character
## Mode :character Mode :character Median :0.0000 Mode :character
## Mean :0.6936
## 3rd Qu.:1.0000
## Max. :3.0000
## furnishingstatus
## Length:545
## Class :character
## Mode :character
##
##
## The dataset contains both numerical and categorical variables, which is well-suited for tree-based models such as Random Forest.
4 Data Cleaning
We perform basic cleaning to ensure data quality:
- remove missing values
- filter out unrealistic area values (e.g., extremely small houses)
# Remove missing values
df <- df %>% drop_na()
# Filter unrealistic area (example rule: > 200 sqft)
df <- df %>% filter(area > 200)
summary(df$area)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1650 3600 4600 5151 6360 162005 Feature Engineering
We engineer additional variables and ensure that categorical fields are correctly encoded as factors.
# Create price per sqft
df <- df %>% mutate(price_per_sqft = price / area)
# Convert categorical variables to factors
df$mainroad <- as.factor(df$mainroad)
df$guestroom <- as.factor(df$guestroom)
df$basement <- as.factor(df$basement)
df$hotwaterheating <- as.factor(df$hotwaterheating)
df$airconditioning <- as.factor(df$airconditioning)
df$prefarea <- as.factor(df$prefarea)
df$furnishingstatus <- as.factor(df$furnishingstatus)
str(df)## 'data.frame': 545 obs. of 14 variables:
## $ price : int 13300000 12250000 12250000 12215000 11410000 10850000 10150000 10150000 9870000 9800000 ...
## $ area : int 7420 8960 9960 7500 7420 7500 8580 16200 8100 5750 ...
## $ bedrooms : int 4 4 3 4 4 3 4 5 4 3 ...
## $ bathrooms : int 2 4 2 2 1 3 3 3 1 2 ...
## $ stories : int 3 4 2 2 2 1 4 2 2 4 ...
## $ mainroad : Factor w/ 2 levels "no","yes": 2 2 2 2 2 2 2 2 2 2 ...
## $ guestroom : Factor w/ 2 levels "no","yes": 1 1 1 1 2 1 1 1 2 2 ...
## $ basement : Factor w/ 2 levels "no","yes": 1 1 2 2 2 2 1 1 2 1 ...
## $ hotwaterheating : Factor w/ 2 levels "no","yes": 1 1 1 1 1 1 1 1 1 1 ...
## $ airconditioning : Factor w/ 2 levels "no","yes": 2 2 1 2 2 2 2 1 2 2 ...
## $ parking : int 2 3 2 3 2 2 2 0 2 1 ...
## $ prefarea : Factor w/ 2 levels "no","yes": 2 1 2 2 1 2 2 1 2 2 ...
## $ furnishingstatus: Factor w/ 3 levels "furnished","semi-furnished",..: 1 1 2 1 1 2 2 3 1 3 ...
## $ price_per_sqft : num 1792 1367 1230 1629 1538 ...6 Exploratory Data Analysis (EDA)
6.1 Price Distribution
ggplot(df, aes(price)) +
geom_density(fill = "#9b59b6", alpha = 0.4) +
labs(
title = "Distribution of House Prices",
x = "Price",
y = "Density"
) +
theme_minimal()
6.2 Bedrooms vs Price
ggplot(df, aes(bedrooms, price)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "lm", color = "#2980b9") +
labs(
title = "Relationship Between Bedrooms and Price",
x = "Number of Bedrooms",
y = "Price"
) +
theme_minimal()
6.3 Area vs Price
ggplot(df, aes(area, price)) +
geom_point(alpha = 0.5) +
geom_smooth(method = "lm", color = "#e74c3c") +
labs(
title = "Area vs Price",
x = "Area",
y = "Price"
) +
theme_minimal()
7 Correlation Analysis
To understand how numerical features are related to price, we compute a correlation matrix.
numeric_df <- df %>%
select(price, area, bedrooms, bathrooms, stories, parking, price_per_sqft)
cor_mat <- cor(numeric_df)
corrplot(cor_mat, method = "color", addCoef.col = "black",
tl.col = "black", number.cex = 0.7,
main = "Correlation Heatmap of Numeric Features")
Insights:
- We expect price to be positively correlated with
area, bedrooms,
bathrooms, and stories.
- Strong correlations highlight which features are likely to be key drivers in the predictive model.
8 Train–Test Split
set.seed(123)
trainIndex <- createDataPartition(df$price, p = 0.8, list = FALSE)
train <- df[trainIndex, ]
test <- df[-trainIndex, ]
nrow(train); nrow(test)## [1] 438## [1] 1079 Modeling
We compare a baseline Linear Regression model with a Random Forest model.
9.1 Linear Regression Model
model_lm <- lm(
price ~ area + bedrooms + bathrooms + stories +
mainroad + guestroom + basement + hotwaterheating +
airconditioning + parking + prefarea + furnishingstatus,
data = train
)
summary(model_lm)##
## Call:
## lm(formula = price ~ area + bedrooms + bathrooms + stories +
## mainroad + guestroom + basement + hotwaterheating + airconditioning +
## parking + prefarea + furnishingstatus, data = train)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2674777 -654183 -41657 482010 4681273
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) -42053.57 287850.03 -0.146 0.883915
## area 252.21 26.46 9.533 < 2e-16 ***
## bedrooms 134851.80 78837.38 1.711 0.087904 .
## bathrooms 976660.21 113788.88 8.583 < 2e-16 ***
## stories 449036.43 69467.54 6.464 2.81e-10 ***
## mainroadyes 464880.56 152408.07 3.050 0.002430 **
## guestroomyes 236678.68 145487.05 1.627 0.104522
## basementyes 417954.12 123001.36 3.398 0.000743 ***
## hotwaterheatingyes 727162.92 230758.29 3.151 0.001741 **
## airconditioningyes 813120.17 119183.43 6.822 3.10e-11 ***
## parking 276313.45 64808.00 4.264 2.48e-05 ***
## prefareayes 577198.71 126844.76 4.550 7.00e-06 ***
## furnishingstatussemi-furnished -91844.86 127520.35 -0.720 0.471776
## furnishingstatusunfurnished -430123.60 139861.29 -3.075 0.002239 **
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1052000 on 424 degrees of freedom
## Multiple R-squared: 0.6804, Adjusted R-squared: 0.6706
## F-statistic: 69.43 on 13 and 424 DF, p-value: < 2.2e-169.2 Random Forest Model
set.seed(123)
model_rf <- randomForest(
price ~ area + bedrooms + bathrooms + stories +
mainroad + guestroom + basement + hotwaterheating +
airconditioning + parking + prefarea + furnishingstatus,
data = train,
ntree = 300,
mtry = 4,
importance = TRUE
)
# Variable importance plot
varImpPlot(model_rf, main = "Random Forest Feature Importance")
10 Model Evaluation
# Predictions on test set
pred_lm <- predict(model_lm, test)
pred_rf <- predict(model_rf, test)
# Compute evaluation metrics
rmse_lm <- rmse(test$price, pred_lm)
mae_lm <- mae(test$price, pred_lm)
rmse_rf <- rmse(test$price, pred_rf)
mae_rf <- mae(test$price, pred_rf)
rmse_lm; mae_lm## [1] 1138053## [1] 834204## [1] 1129612## [1] 820556.410.1 KPI Summary
cat('<div class="kpi-box">',
'<div class="kpi-label">Linear Regression RMSE</div>',
'<div class="kpi-value">', round(rmse_lm, 2), '</div>',
'</div>')
Linear Regression RMSE
1138053
cat('<div class="kpi-box">',
'<div class="kpi-label">Random Forest RMSE</div>',
'<div class="kpi-value">', round(rmse_rf, 2), '</div>',
'</div>')
Random Forest RMSE
1129612
cat('<div class="kpi-box">',
'<div class="kpi-label">Linear Regression MAE</div>',
'<div class="kpi-value">', round(mae_lm, 2), '</div>',
'</div>')
Linear Regression MAE
834204
cat('<div class="kpi-box">',
'<div class="kpi-label">Random Forest MAE</div>',
'<div class="kpi-value">', round(mae_rf, 2), '</div>',
'</div>')
Random Forest MAE
820556.4
10.2 Actual vs Predicted (Random Forest)
comparison_df <- data.frame(
actual = test$price,
predicted = pred_rf
)
ggplot(comparison_df, aes(actual, predicted)) +
geom_point(color = "#27ae60", alpha = 0.6) +
geom_abline(color = "red", linetype = "dashed") +
labs(
title = "Actual vs Predicted Housing Prices (Random Forest)",
x = "Actual Price",
y = "Predicted Price"
) +
theme_minimal()
11 Findings & Business Implications
Based on the analysis and modeling results:
- Size matters: Larger houses (higher
area) and more bedrooms/bathrooms are strongly associated with higher prices.
- Quality and amenities matter: Features such as
air conditioning, basement, and
preferred area (prefarea) contribute positively to
price.
- Parking capacity: Houses with more parking spaces
tend to be priced higher, reflecting convenience and
added value in dense urban settings.
- Furnishing status: Fully or semi-furnished houses often command a price premium over unfurnished properties.
From a real estate industry perspective, these insights can be used to:
- Support fair and consistent pricing when listing
properties.
- Identify high-value renovation opportunities (e.g.,
adding air conditioning or improving furnishing).
- Help buyers understand which attributes drive price differences in
the market.
- Assist banks and valuers in building automated valuation models (AVMs) for mortgage risk assessment.
12 Conclusion & Future Work
This study shows that:
- Data Science techniques can significantly improve housing
price prediction.
- The Random Forest model outperforms Linear
Regression in terms of RMSE and MAE, capturing non-linear relationships
and interactions between features.
- Key drivers of housing price include area, bedrooms, bathrooms, stories, and several amenity-related features.
Future extensions could include:
- Incorporating location-based variables (e.g.,
distance to city center, schools, public transport).
- Using more advanced models (e.g., Gradient Boosting, XGBoost) and
performing hyperparameter tuning.
- Deploying the model as a simple web-based price prediction tool for agents and buyers.