Data Preparation

df <- CleandLaptopData |>

 tibble()

## Due to having too many missing values in the display size, I decided to replace it with average values

mean_dps <- df |>

 select(display_size) |>

 filter(display_size != "Missing") |>

 unlist() |>

 as.numeric()

mean_dps <- as.character(round(mean(mean_dps), 2))

## To select important product specifications that can be used for analyzing

df <- df |>

 select(brand, model, 

     processor_brand, processor_name,

     ram_gb, ram_type, ssd, hdd,

     os, os_bit, graphic_card_gb,

     weight, display_size, Touchscreen,

     latest_price, star_rating) |>

 mutate(display_size = replace(

  display_size, display_size == "Missing", mean_dps))

## Converting to the suitable data types

df$display_size <- as.double(df$display_size) ##double for display_size

df[, -c(13, 15, 16)] <- lapply(df[, -c(13, 15, 16)], as.factor) ##factor for others

The purpose of analyzing this time is to find the preferred components that customers gave a rating more than or equal to 4.0 stars. Instead of simply sorting from the ranking, I chose to sort from 5 brands that have the most product lists in this dataset to get enough information and bring to preferred factors accurately.

## To sort the most numbers of 5 brands in the dataset 

freq_b <- table(df$brand)

sorted_b <- sort(freq_b, decreasing = TRUE)

print(names(sorted_b)[1:5])

## Results

[1] "ASUS"  "DELL"  "Lenovo" "HP"   "acer"

# convert brands to lowercase

df$brand <- as.factor(tolower(df$brand))

## To sort by rating

df <- df[order(df$star_rating, decreasing = TRUE), ]

## To filter only top five frequent brands

top5df <- df |>

 filter(brand %in% c("asus", "dell", "lenovo", "hp", "acer"))

85% of this dataset was recorded by these five brands; Asus, Dell, Lenovo, HP, and Acer accordingly.

## To find processor and ram type in products that get rating >= 4.0

pop_pcsr <- top5df |>

 filter(star_rating >= 4.0) |>

 select(3,4,6) |>

 group_by(processor_brand, processor_name, ram_type) |>

 summarise(n=n(), .groups = "drop") |>

 arrange(desc(n))

Laptops that got more than or equal to 4.0 ratings from customers have processor models including Ryzen 5,7,3, and 9 from AMD. Meanwhile, processors from Intel including Core i5,i3, and i7 that are the most voted processors from filtered data.

## To analyze suitable display size data

## Average and quartile display size

fstar <- top5df |>

 filter(star_rating >= 4.0)

## Average display size

avg_dis <- fstar |>

 summarise(avg_dps = mean(display_size))

## Results of average display size

1  15.2

## Quantile value of display size

q_dis <- fstar |>

    select(display_size)

q_dis <- quantile(q_dis, probs = c(0, .25, .5, .75, 1), na.rm = T)

## Results of all five brands

0%  25%  50%  75% 100% 

13.00 15.12 15.12 15.60 17.30

For the touchscreen function, data reflects that customers do not prioritize that much. Even each laptop does not have a touchscreen for 87.90%, they still got the mentioned rating star.

## Touchscreen ratio

t_dis <- fstar |> count(Touchscreen, sort = T) |>

    mutate(percent = n/sum(n)*100)

This dataset did not specify the currency of all products. For the average price is $67,144.00, meanwhile, the first to the third quartiles are $44,517.25 to $75,997.50 that still got the mentioned rating star.

## Average latest product prices

avg_p <- fstar |>

 summarise(avg_p = mean(latest_price))

## Result for average latest product prices

1 67144

## Quantile value of latest product prices

q_p <- fstar |>

 select(latest_price)

q_p <- quantile(q_p, probs = c(0, .25, .5, .75, 1), na.rm = T)

## Results of all five brands

0%    25%    50%    75%   100% 

19990.00 44517.25 59490.00 75997.50 441990.00