library(stringr)
library(moments)
library(ggplot2)
## Warning: package 'ggplot2' was built under R version 4.5.3
#Loading the data
usedCars<-read.csv("used_cars_dataset.csv", stringsAsFactors = FALSE)
str(usedCars)
## 'data.frame': 14856 obs. of 11 variables:
## $ Brand : chr "Honda" "Toyota" "Volkswagen" "Maruti Suzuki" ...
## $ model : chr "City" "Innova" "VentoTest" "Swift" ...
## $ Year : int 2001 2009 2010 2017 2019 2014 2014 2019 2020 2017 ...
## $ Age : int 23 15 14 7 5 10 10 5 4 7 ...
## $ kmDriven : chr "98,000 km" "190000.0 km" "77,246 km" "83,500 km" ...
## $ Transmission: chr "Manual" "Manual" "Manual" "Manual" ...
## $ Owner : chr "second" "second" "first" "second" ...
## $ FuelType : chr "Petrol" "Diesel" "Diesel" "Diesel" ...
## $ PostedDate : chr "Nov-24" "Jul-24" "Nov-24" "Nov-24" ...
## $ AdditionInfo: chr "Honda City v teck in mint condition, valid genuine car," "Toyota Innova 2.5 G (Diesel) 7 Seater, 2009, Diesel" "Volkswagen Vento 2010-2013 Diesel Breeze, 2010, Diesel" "Maruti Suzuki Swift 2017 Diesel Good Condition" ...
## $ AskPrice : chr "₹ 1,95,000" "₹ 3,75,000" "₹ 1,84,999" "₹ 5,65,000" ...
The dataset initially contained numerical fields such as price and
mileage in text format, which required cleaning and conversion for
analysis.
summary(usedCars)
## Brand model Year Age
## Length:14856 Length:14856 Min. :1900 Min. : 0.000
## Class :character Class :character 1st Qu.:2014 1st Qu.: 5.000
## Mode :character Mode :character Median :2017 Median : 7.000
## Mean :2016 Mean : 7.666
## 3rd Qu.:2019 3rd Qu.: 10.000
## Max. :2024 Max. :124.000
## kmDriven Transmission Owner FuelType
## Length:14856 Length:14856 Length:14856 Length:14856
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
## PostedDate AdditionInfo AskPrice
## Length:14856 Length:14856 Length:14856
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
#Checking missing values
colSums(is.na(usedCars))
## Brand model Year Age kmDriven Transmission
## 0 0 0 0 0 0
## Owner FuelType PostedDate AdditionInfo AskPrice
## 0 0 0 0 0
No missing values were detected, ensuring data completeness and
reliability for further analysis.
#Cleaning the price column
usedCars$AskPrice <- str_replace_all(usedCars$AskPrice, "₹ ","")
usedCars$AskPrice <- str_replace_all(usedCars$AskPrice, ",", "")
usedCars$AskPrice <- as.numeric(usedCars$AskPrice)
str(usedCars)
## 'data.frame': 14856 obs. of 11 variables:
## $ Brand : chr "Honda" "Toyota" "Volkswagen" "Maruti Suzuki" ...
## $ model : chr "City" "Innova" "VentoTest" "Swift" ...
## $ Year : int 2001 2009 2010 2017 2019 2014 2014 2019 2020 2017 ...
## $ Age : int 23 15 14 7 5 10 10 5 4 7 ...
## $ kmDriven : chr "98,000 km" "190000.0 km" "77,246 km" "83,500 km" ...
## $ Transmission: chr "Manual" "Manual" "Manual" "Manual" ...
## $ Owner : chr "second" "second" "first" "second" ...
## $ FuelType : chr "Petrol" "Diesel" "Diesel" "Diesel" ...
## $ PostedDate : chr "Nov-24" "Jul-24" "Nov-24" "Nov-24" ...
## $ AdditionInfo: chr "Honda City v teck in mint condition, valid genuine car," "Toyota Innova 2.5 G (Diesel) 7 Seater, 2009, Diesel" "Volkswagen Vento 2010-2013 Diesel Breeze, 2010, Diesel" "Maruti Suzuki Swift 2017 Diesel Good Condition" ...
## $ AskPrice : num 195000 375000 184999 565000 685000 ...
The mileage column was standardized by removing text and converting
it into numeric format for accurate analysis.
#Standardizing the text data
usedCars$Transmission <- str_to_title(usedCars$Transmission)
usedCars$Owner <- str_to_title(usedCars$Owner)
usedCars$FuelType <- str_to_title(usedCars$FuelType)
Text variables were standardized (converted to title case) to avoid
inconsistencies caused by different capitalizations.
# Removing duplicates
usedCars <- unique(usedCars)
str(usedCars)
## 'data.frame': 13850 obs. of 11 variables:
## $ Brand : chr "Honda" "Toyota" "Volkswagen" "Maruti Suzuki" ...
## $ model : chr "City" "Innova" "VentoTest" "Swift" ...
## $ Year : int 2001 2009 2010 2017 2019 2014 2014 2019 2020 2017 ...
## $ Age : int 23 15 14 7 5 10 10 5 4 7 ...
## $ kmDriven : num 98000 190000 77246 83500 45000 ...
## $ Transmission: chr "Manual" "Manual" "Manual" "Manual" ...
## $ Owner : chr "Second" "Second" "First" "Second" ...
## $ FuelType : chr "Petrol" "Diesel" "Diesel" "Diesel" ...
## $ PostedDate : chr "Nov-24" "Jul-24" "Nov-24" "Nov-24" ...
## $ AdditionInfo: chr "Honda City v teck in mint condition, valid genuine car," "Toyota Innova 2.5 G (Diesel) 7 Seater, 2009, Diesel" "Volkswagen Vento 2010-2013 Diesel Breeze, 2010, Diesel" "Maruti Suzuki Swift 2017 Diesel Good Condition" ...
## $ AskPrice : num 195000 375000 184999 565000 685000 ...
Duplicate records were removed to improve data quality and prevent
biased analysis.
# Handling outliers
boxplot(usedCars$kmDriven)
usedCars <- usedCars[usedCars$kmDriven > 1000 & usedCars$kmDriven < 150000, ]
# Handling outliers
boxplot(usedCars$Year)
usedCars <- usedCars[usedCars$Year > 1999, ]
usedCars <- usedCars[usedCars$AskPrice < 15000000, ]
Outliers were filtered to ensure realistic data ranges and improve
model accuracy.
#Saving cleaned dataset
write.csv(usedCars, "used_cars_clean.csv", row.names = FALSE)
#EDA
#1. Overall summary, mean, median, mode, standard deviation
summary(usedCars)
## Brand model Year Age
## Length:13050 Length:13050 Min. :2000 Min. : 0.000
## Class :character Class :character 1st Qu.:2014 1st Qu.: 5.000
## Mode :character Mode :character Median :2017 Median : 7.000
## Mean :2016 Mean : 7.524
## 3rd Qu.:2019 3rd Qu.:10.000
## Max. :2024 Max. :24.000
## kmDriven Transmission Owner FuelType
## Min. : 1022 Length:13050 Length:13050 Length:13050
## 1st Qu.: 44330 Class :character Class :character Class :character
## Median : 65000 Mode :character Mode :character Mode :character
## Mean : 65047
## 3rd Qu.: 84515
## Max. :149500
## PostedDate AdditionInfo AskPrice
## Length:13050 Length:13050 Min. : 15000
## Class :character Class :character 1st Qu.: 350000
## Mode :character Mode :character Median : 570000
## Mean : 960956
## 3rd Qu.: 995000
## Max. :14900000
The dataset shows a wide range of values across price, mileage, and
manufacturing year. The spread indicates a diverse mix of vehicles from
different segments, ranging from low-cost older cars to relatively newer
and higher-priced vehicles.
mean(usedCars$AskPrice)
## [1] 960956.2
median(usedCars$AskPrice)
## [1] 570000
sd(usedCars$AskPrice)
## [1] 1293931
The interquartile values show that 50% of the cars lie within a
moderate price range. This confirms that most vehicles fall in the
mid-market segment, with fewer cars in extreme high or low price
ranges.
#3. Skewness
skewness(usedCars$AskPrice)
## [1] 4.451038
The skewness value is positive, means that the price distribution is
right-skewed, confirming the presence of a few high-priced cars
influencing the distribution.
#4
range(usedCars$AskPrice)
## [1] 15000 14900000
Price range shows a large gap between minimum and maximum values,
indicating a wide market spectrum.
range(usedCars$kmDriven)
## [1] 1022 149500
kmDriven varies significantly, suggesting different levels of
vehicle usage.
range(usedCars$Year)
## [1] 2000 2024
Year range confirms presence of both older and relatively newer
vehicles.
#5. Inter Quartile Range
IQR(usedCars$AskPrice)
## [1] 645000
IQR(usedCars$kmDriven)
## [1] 40184.75
The IQR values indicate the spread of the middle 50% of the data. A
relatively high IQR for price suggests variability even within the
central portion of the dataset, highlighting market diversity.
#6. Correlation
cor(usedCars$AskPrice, usedCars$kmDriven) # negative relation- the more kms driven, lesser the price
## [1] -0.2403548
cor(usedCars$AskPrice, usedCars$Year) # positive relation- latest year, higher price
## [1] 0.3497914
Price decreases with higher mileage and increases with newer
manufacturing year, confirming expected depreciation trends.
#7. frequency table
table(usedCars$FuelType)
##
## Diesel Hybrid Hybrid/Cng Petrol
## 4378 1735 1699 5238
Petrol cars dominate the dataset, followed by diesel
table(usedCars$Transmission)
##
## Automatic Manual
## 6462 6588
Manual transmission vehicles are slightly more common than
automatic
#8. proportion
prop.table(table(usedCars$Brand))
##
## Ambassador Audi Bajaj BMW Chevrolet
## 7.662835e-05 2.084291e-02 3.065134e-04 2.842912e-02 8.735632e-03
## Citroen Datsun Fiat Force Ford
## 1.532567e-04 3.524904e-03 1.992337e-03 3.831418e-04 2.758621e-02
## Honda Hummer Hyundai Isuzu Jaguar
## 8.865900e-02 7.662835e-05 1.704981e-01 8.429119e-04 3.601533e-03
## Jeep Kia Land Rover Lexus Mahindra
## 7.432950e-03 1.578544e-02 6.973180e-03 1.992337e-03 4.996169e-02
## Maruti Suzuki Maserati Mercedes-Benz MG Mini
## 3.115709e-01 7.662835e-05 3.954023e-02 8.122605e-03 3.065134e-03
## Mitsubishi Nissan Opel Porsche Renault
## 1.379310e-03 8.505747e-03 1.532567e-04 1.992337e-03 2.643678e-02
## Skoda Ssangyong Tata Toyota Toyota Land
## 1.938697e-02 3.831418e-04 4.038314e-02 6.337165e-02 1.532567e-04
## Volkswagen Volvo
## 3.310345e-02 4.521073e-03
A few brands dominate the market, indicating brand concentration in
the used car segment.
prop.table(table(usedCars$FuelType))
##
## Diesel Hybrid Hybrid/Cng Petrol
## 0.3354789 0.1329502 0.1301916 0.4013793
Diesel cars have a higher average resale value compared to petrol
and other fuel types. This may be due to better fuel efficiency and
higher demand for long-distance usage.
aggregate(AskPrice ~ Brand, usedCars, mean)
## Brand AskPrice
## 1 Ambassador 421000.0
## 2 Audi 1850168.9
## 3 Bajaj 219750.0
## 4 BMW 2917848.1
## 5 Chevrolet 246030.2
## 6 Citroen 1950000.0
## 7 Datsun 240826.0
## 8 Fiat 255957.7
## 9 Force 1252200.0
## 10 Ford 1030233.0
## 11 Honda 510758.3
## 12 Hummer 9800000.0
## 13 Hyundai 578536.4
## 14 Isuzu 1696272.7
## 15 Jaguar 2614553.1
## 16 Jeep 1758600.3
## 17 Kia 1466524.7
## 18 Land Rover 5133880.3
## 19 Lexus 4851923.0
## 20 Mahindra 961130.5
## 21 Maruti Suzuki 485533.8
## 22 Maserati 8000000.0
## 23 Mercedes-Benz 3098073.0
## 24 MG 1605113.3
## 25 Mini 2921975.0
## 26 Mitsubishi 777344.4
## 27 Nissan 442872.1
## 28 Opel 129000.0
## 29 Porsche 7371884.6
## 30 Renault 439164.0
## 31 Skoda 960030.6
## 32 Ssangyong 508000.6
## 33 Tata 670779.5
## 34 Toyota 1775640.0
## 35 Toyota Land 5700000.0
## 36 Volkswagen 670246.2
## 37 Volvo 2739775.8
Certain brands command higher average prices, indicating stronger
brand value and perceived reliability in the resale market.
#10. median price by fuel
aggregate(AskPrice ~ FuelType, data = usedCars, median)
## FuelType AskPrice
## 1 Diesel 899999.5
## 2 Hybrid 410000.0
## 3 Hybrid/Cng 425000.0
## 4 Petrol 515000.0
Automatic transmission vehicles have higher average prices compared
to manual ones. This reflects growing consumer preference for
convenience.
#12. avg price by owner type
aggregate(AskPrice ~ Owner, data = usedCars, mean)
## Owner AskPrice
## 1 First 1147832.4
## 2 Second 776752.9
First-owner vehicles retain higher value, while resale price
decreases with additional ownership.
#13. fuel vs kmdriven
aggregate(kmDriven ~ FuelType, usedCars, mean)
## FuelType kmDriven
## 1 Diesel 76463.51
## 2 Hybrid 67121.10
## 3 Hybrid/Cng 66378.24
## 4 Petrol 54386.93
Diesel cars are driven more on average.
#14. multi-dimensional analysis-fuel+transmission
aggregate(AskPrice ~ FuelType + Transmission, data = usedCars, mean)
## FuelType Transmission AskPrice
## 1 Diesel Automatic 2039212.1
## 2 Hybrid Automatic 617058.8
## 3 Hybrid/Cng Automatic 674619.5
## 4 Petrol Automatic 1320305.8
## 5 Diesel Manual 708977.6
## 6 Hybrid Manual 445974.3
## 7 Hybrid/Cng Manual 441661.4
## 8 Petrol Manual 447311.0
Diesel + Automatic combinations tend to have the highest prices,
indicating a premium segment combination.
Petrol + Manual combinations dominate lower price ranges
#15. age impact on price
cor(usedCars$AskPrice, usedCars$Age)
## [1] -0.3497914
There is a negative correlation between Age and Price. Older cars
are priced lower due to depreciation
#16. grouping by age buckets
usedCars$Age_Group <- cut(usedCars$Age,
breaks = c(0,3,7,15),
labels = c("New", "Mid", "Old"))
aggregate(AskPrice ~ Age_Group, data = usedCars, mean)
## Age_Group AskPrice
## 1 New 1703621.3
## 2 Mid 1122099.9
## 3 Old 580364.6
New cars are priced highest, followed by mid-age and old cars,
confirming depreciation patterns.
#17. segmenting by price range
usedCars$Price_Category <- cut(usedCars$AskPrice,
breaks = c(0,300000, 700000, 2000000),
labels = c("Low", "Medium", "High"))
#counting by segment
table(usedCars$Price_Category)
##
## Low Medium High
## 2650 5462 3679
Most cars fall in the medium price segment, indicating a balanced
market with fewer luxury vehicles.
#18. transmission by segment
table(usedCars$Transmission, usedCars$Price_Category)
##
## Low Medium High
## Automatic 741 2180 2330
## Manual 1909 3282 1349
# owner by segment
table(usedCars$Owner, usedCars$Price_Category)
##
## Low Medium High
## First 731 2696 2251
## Second 1919 2766 1428
Premium segments are dominated by automatic and first-owner
vehicles.
#19. fuel distribution in each segment
table(usedCars$FuelType, usedCars$Price_Category)
##
## Low Medium High
## Diesel 314 1317 1900
## Hybrid 580 872 245
## Hybrid/Cng 536 887 233
## Petrol 1220 2386 1301
Diesel cars are more common in high-price segments, while petrol
dominates lower segments.
#20. coefficient of variation- measuring price volatility
sd(usedCars$AskPrice)/mean(usedCars$AskPrice)
## [1] 1.346504
High variation in price indicates a heterogeneous market with
diverse pricing structures.
#21. most expensive cars
usedCars[which.max(usedCars$AskPrice), ]
## Brand model Year Age kmDriven Transmission Owner FuelType
## 13487 Porsche Cayenne Coupe 2022 2 9000 Automatic First Petrol
## PostedDate AdditionInfo AskPrice Age_Group
## 13487 Nov-24 Porsche Cayenne Coupe, 2022, Petrol 14900000 New
## Price_Category
## 13487 <NA>
# top 5 expensive cars-
head(usedCars[order(-usedCars$AskPrice), ], 5)
## Brand model Year Age kmDriven Transmission Owner
## 13487 Porsche Cayenne Coupe 2022 2 9000 Automatic First
## 1022 Mercedes-Benz G-Class 2017 7 45000 Automatic Second
## 6986 Land Rover Range Rover Sport 2020 4 30000 Automatic First
## 7130 Land Rover Range Rover 2019 5 110000 Automatic Second
## 6028 Mercedes-Benz S-Class 2023 1 22000 Automatic First
## FuelType PostedDate
## 13487 Petrol Nov-24
## 1022 Petrol Nov-24
## 6986 Petrol Sep-24
## 7130 Diesel Nov-24
## 6028 Diesel Nov-24
## AdditionInfo AskPrice
## 13487 Porsche Cayenne Coupe, 2022, Petrol 14900000
## 1022 Mercedes-Benz G-Class AMG G 63 Grand Edition, 2017, Petrol 14500000
## 6986 Land Rover Range Sport HSE, 2020, Petrol 14500000
## 7130 Land Rover Range 4.4 Diesel LWB SVAutobiography, 2019, Diesel 14000000
## 6028 Mercedes-Benz S-Class S 350d, 2023, Diesel 13800000
## Age_Group Price_Category
## 13487 New <NA>
## 1022 Mid <NA>
## 6986 Mid <NA>
## 7130 Mid <NA>
## 6028 New <NA>
The most expensive car represents a premium segment vehicle with
likely low mileage, recent manufacturing year, and desirable
features.
# 5 cheapest cars
head(usedCars[order(usedCars$AskPrice), ], 5)
## Brand model Year Age kmDriven Transmission Owner FuelType
## 1736 Mahindra Marshal 2000 24 55000 Manual Second Diesel
## 13470 BMW 5 Series 2014 10 90000 Automatic Second Diesel
## 14315 Renault Duster 2013 11 80000 Manual Second Diesel
## 3014 Maruti Suzuki Baleno 2024 0 9000 Manual First Petrol
## 270 Maruti Suzuki Alto-800 2006 18 20024 Manual First Petrol
## PostedDate AdditionInfo AskPrice Age_Group
## 1736 Nov-24 Want NOC for Mahindra Marshal or Commander 15000 <NA>
## 13470 Nov-24 BMW 5 Series 2014 15000 Old
## 14315 Nov-24 2013 duster 110 all dpare available 15000 Old
## 3014 Nov-24 Panasound brand Daimond 2k androd stereo 18500 <NA>
## 270 Nov-24 पुरानी स्क्रैप की गाड़ी बेचने के लिए संपर्क करें 20000 <NA>
## Price_Category
## 1736 Low
## 13470 Low
## 14315 Low
## 3014 Low
## 270 Low
#22. avg price by segment
aggregate(AskPrice ~ Price_Category, data = usedCars, mean)
## Price_Category AskPrice
## 1 Low 204319.4
## 2 Medium 492294.7
## 3 High 1148201.7
Clear separation exists between low, medium, and high segments,
validating the segmentation approach.
#23. which combo gives the highest price
aggregate(AskPrice ~ FuelType + Transmission, data = usedCars, max)
## FuelType Transmission AskPrice
## 1 Diesel Automatic 14000000
## 2 Hybrid Automatic 11000000
## 3 Hybrid/Cng Automatic 11900000
## 4 Petrol Automatic 14900000
## 5 Diesel Manual 4550000
## 6 Hybrid Manual 1450000
## 7 Hybrid/Cng Manual 1550000
## 8 Petrol Manual 6500000
The highest priced vehicles are typically diesel automatic cars,
indicating strong demand for this combination in the resale market.
library(ggplot2)
ggplot(usedCars, aes(x=AskPrice/100000)) +
geom_histogram(bins=15) +
labs(x = "Price (in Lakhs)", y="Number of cars")
The price distribution is right-skewed, with most cars concentrated
in lower price ranges.
# Relationship Analysis
# Price vs kmDriven (Regression)
ggplot(usedCars, aes(x=kmDriven, y=AskPrice/100000)) +
geom_point() +
geom_smooth(method="lm") +
labs(y= "Price (in Lakhs)", x="Number of Kilometers")
## `geom_smooth()` using formula = 'y ~ x'
A negative relationship is observed, where higher mileage leads to
lower resale value.
# price vs year - Depreciation Trend
ggplot(usedCars, aes(x=Year, y=AskPrice/100000)) + geom_point() + labs(y = "Price in lakhs", x="Year")
Newer cars command higher prices, indicating depreciation over
time.
# Price by Transmission (Boxplot)
ggplot(usedCars, aes(x=Transmission, y=AskPrice/100000)) + geom_boxplot() + labs(y = "Price in lakhs")
# Price by Fuel (Boxplot)
ggplot(usedCars, aes(x=FuelType, y=AskPrice/100000)) + geom_boxplot() + labs(y="Price in lakhs")
# Price by owner
ggplot(usedCars, aes(x=Owner, y=AskPrice/100000)) + geom_boxplot() + labs(y="Price in lakhs")
Price varies significantly across fuel types, transmission, and
ownership categories.
library(corrplot)
## Warning: package 'corrplot' was built under R version 4.5.3
## corrplot 0.95 loaded
corrplot(cor(usedCars[sapply(usedCars, is.numeric)]))
The correlation plot shows that price increases with newer
manufacturing year and decreases with higher age and mileage. It
confirms that age and usage are the primary factors affecting car
prices.
# Price vs kmDriven by Fuel
ggplot(usedCars, aes(kmDriven, (AskPrice/100000), color=FuelType)) + geom_point() + labs(y="Price in lakhs", x="Kms Driven")
Price decreases as km driven increases, showing depreciation. Diesel
cars retain higher value, while petrol and hybrid cars are mostly in
lower price ranges.
# faceted scatter plot by fuel
ggplot(usedCars, aes(kmDriven, AskPrice/100000)) + geom_point() + facet_wrap(~FuelType) + labs(y="Price in lakhs")
Price decreases with mileage across all fuel types. Diesel cars
retain higher value, while petrol and hybrid/CNG cars are mostly in
lower price ranges.
# Building a linear relationship model
# Convert categorical variables
usedCars$FuelType <- as.factor(usedCars$FuelType)
usedCars$Transmission <- as.factor(usedCars$Transmission)
usedCars$Owner <- as.factor(usedCars$Owner)
usedCars$Brand <- as.factor(usedCars$Brand)
# TRAIN-TEST SPLIT (70–30)
set.seed(123)
train_index <- sample(1:nrow(usedCars), 0.7 * nrow(usedCars))
train <- usedCars[train_index, ]
test <- usedCars[-train_index, ]
model <- lm(AskPrice ~ Age + kmDriven + FuelType + Transmission + Owner, data=train)
summary(model)
##
## Call:
## lm(formula = AskPrice ~ Age + kmDriven + FuelType + Transmission +
## Owner, data = train)
##
## Residuals:
## Min 1Q Median 3Q Max
## -2153555 -506765 -169664 214052 12981638
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 2.773e+06 3.738e+04 74.19 <2e-16 ***
## Age -7.389e+04 3.486e+03 -21.20 <2e-16 ***
## kmDriven -6.526e+00 4.650e-01 -14.03 <2e-16 ***
## FuelTypeHybrid -8.725e+05 3.679e+04 -23.71 <2e-16 ***
## FuelTypeHybrid/Cng -8.813e+05 3.727e+04 -23.64 <2e-16 ***
## FuelTypePetrol -6.483e+05 2.848e+04 -22.76 <2e-16 ***
## TransmissionManual -6.918e+05 2.298e+04 -30.11 <2e-16 ***
## OwnerSecond -4.969e+03 2.490e+04 -0.20 0.842
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## Residual standard error: 1082000 on 9127 degrees of freedom
## Multiple R-squared: 0.2806, Adjusted R-squared: 0.28
## F-statistic: 508.5 on 7 and 9127 DF, p-value: < 2.2e-16
The model shows that age, mileage, fuel type, and transmission
significantly influence price. However, the model explains only ~28% of
variation, indicating limited predictive power.
pred <- predict(model, newdata=test)
The Random Forest model captures non-linear relationships and
improves prediction accuracy, explaining ~74% of variation in
prices.
importance(rf_model)
## IncNodePurity
## Age 3.194804e+14
## kmDriven 1.109744e+14
## FuelType 1.573757e+14
## Transmission 9.263559e+13
## Owner 2.190703e+13
## Brand 4.397035e+14
varImpPlot(rf_model)
Brand, age, and mileage are the most important predictors of
price.
Final Conclusion: The analysis demonstrates that used car prices are
influenced by multiple factors, with age, mileage, brand, and fuel type
being the most significant. While linear regression provides basic
insights, Random Forest delivers superior predictive performance by
capturing complex relationships in the data.