Netflix
Surya Teja Madhunala
2025-12-04
Issue Description
In recent years, streaming platforms have become a dominant way for
people to watch films and TV shows. Netflix periodically publishes
engagement reports summarizing what people watched and how many hours
they spent watching different titles.
In this project, I analyze Netflix’s viewing data from late 2023
through the first half of 2025 using the TidyTuesday “Netflix viewing”
dataset. I focus on understanding which kinds of titles attract the most
viewing time, how movies and shows differ in engagement, and whether
runtime is related to total viewing.
Questions
Do TV shows or movies account for more total hours viewed on
Netflix?
Which titles and genres dominate viewing time during the period
covered by the dataset?
Is there a relationship between runtime and engagement (hours
viewed and “views” = hours viewed/runtime)?
Are newer titles watched more than older ones, or do older
catalog titles still generate substantial viewing?
Data Source
I use two CSV files: 1. movies.csv – viewing statistics for movies
Url: https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-07-29/movies.csv
2. shows.csv – viewing statistics for series “https://raw.githubusercontent.com/rfordatascience/tidytuesday/main/data/2025/2025-07-29/shows.csv”
# Load packages
library(tidytuesdayR)
## Warning: package 'tidytuesdayR' was built under R version 4.5.2
## Warning: package 'readr' was built under R version 4.5.2
## Warning: package 'dplyr' was built under R version 4.5.2
##
## Attaching package: 'dplyr'
## The following objects are masked from 'package:stats':
##
## filter, lag
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union
## Warning: package 'lubridate' was built under R version 4.5.2
##
## Attaching package: 'lubridate'
## The following objects are masked from 'package:base':
##
## date, intersect, setdiff, union
## Warning: package 'stringr' was built under R version 4.5.2
## Warning: package 'ggplot2' was built under R version 4.5.2
#Load Datasets
# Option 1: tidytuesdayR
tuesdata <- tidytuesdayR::tt_load('2025-07-29')
## ---- Compiling #TidyTuesday Information for 2025-07-29 ----
## --- There are 2 files available ---
##
##
## ── Downloading files ───────────────────────────────────────────────────────────
##
## 1 of 2: "movies.csv"
## 2 of 2: "shows.csv"
movies <- tuesdata$movies
shows <- tuesdata$shows
Documentation
The dataset is documented on the TidyTuesday GitHub repository and in
accompanying blog posts that describe how the Netflix engagement reports
were converted into tidy data.
Each file includes a simple data dictionary: movies.csv / shows.csv
key variables:
source – original file/report name
report – reporting period label (e.g. "2025Jan-Jun")
title – movie or show title
available_globally – "Yes" / "No"
release_date – date when the title was released on Netflix
hours_viewed – total hours viewed during the report period
runtime – runtime stored as a Period (e.g. 1H 54M 0S)
views – hours_viewed / runtime (approximate number of complete views)
Description of the Data
- I first inspect the structure and basic summaries of the two
datasets.
## spc_tbl_ [36,121 × 8] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
## $ source : chr [1:36121] "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" ...
## $ report : chr [1:36121] "2025Jan-Jun" "2025Jan-Jun" "2025Jan-Jun" "2025Jan-Jun" ...
## $ title : chr [1:36121] "Back in Action" "STRAW" "The Life List" "Exterritorial" ...
## $ available_globally: chr [1:36121] "Yes" "Yes" "Yes" "Yes" ...
## $ release_date : Date[1:36121], format: "2025-01-17" "2025-06-06" ...
## $ hours_viewed : num [1:36121] 3.13e+08 1.85e+08 1.99e+08 1.59e+08 1.55e+08 ...
## $ runtime : chr [1:36121] "1H 54M 0S" "1H 48M 0S" "2H 5M 0S" "1H 49M 0S" ...
## $ views : num [1:36121] 1.65e+08 1.03e+08 9.55e+07 8.75e+07 8.69e+07 ...
## - attr(*, "spec")=
## .. cols(
## .. source = col_character(),
## .. report = col_character(),
## .. title = col_character(),
## .. available_globally = col_character(),
## .. release_date = col_date(format = ""),
## .. hours_viewed = col_double(),
## .. runtime = col_character(),
## .. views = col_double()
## .. )
## - attr(*, "problems")=<externalptr>
## source report title available_globally
## Length:36121 Length:36121 Length:36121 Length:36121
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## release_date hours_viewed runtime
## Min. :2013-12-12 Min. : 100000 Length:36121
## 1st Qu.:2019-11-26 1st Qu.: 200000 Class :character
## Median :2021-08-13 Median : 400000 Mode :character
## Mean :2021-05-19 Mean : 2790858
## 3rd Qu.:2022-12-15 3rd Qu.: 1900000
## Max. :2025-06-30 Max. :313000000
## NA's :29396 NA's :12
## views
## Min. : 100000
## 1st Qu.: 100000
## Median : 300000
## Mean : 1573256
## 3rd Qu.: 1100000
## Max. :164700000
## NA's :12
## spc_tbl_ [27,803 × 8] (S3: spec_tbl_df/tbl_df/tbl/data.frame)
## $ source : chr [1:27803] "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" ...
## $ report : chr [1:27803] "2025Jan-Jun" "2025Jan-Jun" "2025Jan-Jun" "2025Jan-Jun" ...
## $ title : chr [1:27803] "Adolescence: Limited Series" "Squid Game: Season 2 // 오징어 게임: 시즌 2" "Squid Game: Season 3 // 오징어 게임: 시즌 3" "Zero Day: Limited Series" ...
## $ available_globally: chr [1:27803] "Yes" "Yes" "Yes" "Yes" ...
## $ release_date : Date[1:27803], format: "2025-03-13" "2024-12-26" ...
## $ hours_viewed : num [1:27803] 5.55e+08 8.40e+08 4.39e+08 3.16e+08 2.19e+08 ...
## $ runtime : chr [1:27803] "3H 50M 0S" "7H 10M 0S" "6H 8M 0S" "5H 9M 0S" ...
## $ views : num [1:27803] 1.45e+08 1.17e+08 7.15e+07 6.13e+07 5.80e+07 ...
## - attr(*, "spec")=
## .. cols(
## .. source = col_character(),
## .. report = col_character(),
## .. title = col_character(),
## .. available_globally = col_character(),
## .. release_date = col_date(format = ""),
## .. hours_viewed = col_double(),
## .. runtime = col_character(),
## .. views = col_double()
## .. )
## - attr(*, "problems")=<externalptr>
## source report title available_globally
## Length:27803 Length:27803 Length:27803 Length:27803
## Class :character Class :character Class :character Class :character
## Mode :character Mode :character Mode :character Mode :character
##
##
##
##
## release_date hours_viewed runtime
## Min. :2010-04-01 Min. : 100000 Length:27803
## 1st Qu.:2019-04-17 1st Qu.: 800000 Class :character
## Median :2021-04-14 Median : 2500000 Mode :character
## Mean :2021-01-08 Mean : 9807243
## 3rd Qu.:2022-12-16 3rd Qu.: 8200000
## Max. :2025-06-30 Max. :840300000
## NA's :14033 NA's :12
## views
## Min. : 100000
## 1st Qu.: 200000
## Median : 400000
## Mean : 1414994
## 3rd Qu.: 1300000
## Max. :144800000
## NA's :12
- To compare movies and shows together, I create a combined dataset
with a type column:
viewing <- bind_rows(
movies |> mutate(type = "Movie"),
shows |> mutate(type = "Show")
)
str(viewing)
## tibble [63,924 × 9] (S3: tbl_df/tbl/data.frame)
## $ source : chr [1:63924] "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" "1_What_We_Watched_A_Netflix_Engagement_Report_2025Jan-Jun" ...
## $ report : chr [1:63924] "2025Jan-Jun" "2025Jan-Jun" "2025Jan-Jun" "2025Jan-Jun" ...
## $ title : chr [1:63924] "Back in Action" "STRAW" "The Life List" "Exterritorial" ...
## $ available_globally: chr [1:63924] "Yes" "Yes" "Yes" "Yes" ...
## $ release_date : Date[1:63924], format: "2025-01-17" "2025-06-06" ...
## $ hours_viewed : num [1:63924] 3.13e+08 1.85e+08 1.99e+08 1.59e+08 1.55e+08 ...
## $ runtime : chr [1:63924] "1H 54M 0S" "1H 48M 0S" "2H 5M 0S" "1H 49M 0S" ...
## $ views : num [1:63924] 1.65e+08 1.03e+08 9.55e+07 8.75e+07 8.69e+07 ...
## $ type : chr [1:63924] "Movie" "Movie" "Movie" "Movie" ...
summary(viewing$hours_viewed)
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## 100000 300000 1000000 5842382 4200000 840300000 24
## Min. 1st Qu. Median Mean 3rd Qu. Max. NA's
## 100000 100000 300000 1504426 1200000 164700000 24
Cleaning and Preparation
Main cleaning steps:
Combine movies and shows into one data frame with a type
indicator.
Convert runtime to minutes so that it is numeric.
Create derived variables such as release year.
Filter out rows with missing or obviously problematic
values.
parse_runtime <- function(x) {
hours <- str_extract(x, "(\\d+)H")
mins <- str_extract(x, "(\\d+)M")
secs <- str_extract(x, "(\\d+)S")
hours <- ifelse(is.na(hours), 0, as.numeric(str_remove(hours, "H")))
mins <- ifelse(is.na(mins), 0, as.numeric(str_remove(mins, "M")))
secs <- ifelse(is.na(secs), 0, as.numeric(str_remove(secs, "S")))
total_minutes <- hours*60 + mins + secs/60
return(total_minutes)
}
viewing_clean <- viewing |>
mutate(
runtime_minutes = parse_runtime(runtime),
runtime_hours = runtime_minutes / 60,
release_year = year(release_date)
) |>
filter(
!is.na(hours_viewed),
!is.na(views),
!is.na(runtime_minutes),
runtime_minutes > 0
)
summary(viewing_clean$runtime_minutes)
## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 2.0 99.0 119.0 260.4 289.0 23853.0
Final Results
Q1 – Do TV shows or movies account for more total hours viewed?
hours_by_type <- viewing_clean |>
group_by(type) |>
summarise(
total_hours = sum(hours_viewed, na.rm = TRUE),
median_hours = median(hours_viewed, na.rm = TRUE),
n_titles = n()
)
hours_by_type
## # A tibble: 2 × 4
## type total_hours median_hours n_titles
## <chr> <dbl> <dbl> <int>
## 1 Movie 100775100000 400000 36109
## 2 Show 262387100000 2400000 27089
ggplot(hours_by_type, aes(x = type, y = total_hours)) +
geom_col() +
labs(
title = "Total Netflix Hours Viewed by Type",
x = "Type",
y = "Total hours viewed (reporting period)"
)
Q2 – Which titles dominate viewing time?
# Top 10 most viewed titles (movies + shows)
top_titles <- viewing_clean |>
arrange(desc(hours_viewed)) |>
slice_head(n = 10)
top_titles
## # A tibble: 10 × 12
## source report title available_globally release_date hours_viewed runtime
## <chr> <chr> <chr> <chr> <date> <dbl> <chr>
## 1 1_What_We_… 2025J… Squi… Yes 2024-12-26 840300000 7H 10M…
## 2 3_What_We_… 2024J… Brid… Yes 2024-05-16 733800000 7H 59M…
## 3 3_What_We_… 2024J… Fool… Yes 2024-01-01 689500000 6H 25M…
## 4 3_What_We_… 2024J… Quee… Yes 2024-03-09 682600000 23H 22…
## 5 4_What_We_… 2023J… King… Yes 2023-06-17 630200000 19H 0M…
## 6 2_What_We_… 2024J… Squi… Yes 2024-12-26 619900000 7H 10M…
## 7 1_What_We_… 2025J… When… Yes 2025-03-07 577000000 16H 35…
## 8 1_What_We_… 2025J… Adol… Yes 2025-03-13 555100000 3H 50M…
## 9 2_What_We_… 2024J… Mons… Yes 2024-09-19 550400000 7H 54M…
## 10 4_What_We_… 2023J… ONE … Yes 2023-08-31 541900000 7H 34M…
## # ℹ 5 more variables: views <dbl>, type <chr>, runtime_minutes <dbl>,
## # runtime_hours <dbl>, release_year <dbl>
ggplot(top_titles,
aes(x = reorder(title, hours_viewed),
y = hours_viewed,
fill = type)) +
geom_col() +
coord_flip() +
labs(
title = "Top 10 Netflix Titles by Hours Viewed",
x = "Title",
y = "Hours Viewed"
) +
theme(legend.position = "bottom")
ggplot(top_titles,
aes(x = reorder(title, hours_viewed),
y = hours_viewed,
fill = type)) +
geom_col() +
coord_flip() +
labs(
title = "Top 10 Netflix Titles by Hours Viewed",
x = "Title",
y = "Hours viewed"
) +
theme(legend.position = "bottom")
Q4 – Are newer titles watched more than older ones?
ggplot(viewing_clean,
aes(x = release_year,
y = hours_viewed,
color = type)) +
geom_point(alpha = 0.4, position = position_jitter(width = 0.2)) +
geom_smooth(se = FALSE) +
labs(
title = "Hours Viewed vs. Release Year",
x = "Release year",
y = "Hours viewed"
) +
theme(legend.position = "bottom")
## `geom_smooth()` using method = 'gam' and formula = 'y ~ s(x, bs = "cs")'
## Warning: Removed 42804 rows containing non-finite outside the scale range
## (`stat_smooth()`).
## Warning: Removed 42804 rows containing missing values or values outside the scale range
## (`geom_point()`).
