Introduction

This report will use data visualization to establish any relationships there may be within this data set of the show “The Office.” Specifically, it will highlight how the show’s popularity and its appeal has grown over the course of the series, whether those two things moved together, and what a handful of unusual episodes can tell us about the limits of that relationship. The data set, office_ratings, contains 186 observations (one per episode for all episodes of the show) and 7 variables, combining U.S. Nielsen viewership figures with IMDb community ratings. It was compiled from publicly available Office ratings and IMDb data.

The full data set can be viewed below:

Distributions of Viewers, IMDB Ratings, and Total Votes

ggplot(office_ratings, aes(x = viewers)) +
  geom_histogram(binwidth = 1) +
  labs(title = "Distribution of Episode Viewership",
    x = "Viewers (millions)",
    y = "Number of Episodes")

The viewership distribution is right-skewed, with most episodes clustered between about 5 and 9 million viewers with its peak being around 7–8 million. The episode, “Stress Relief” (which aired right after the Super Bowl), pulls the tail far to the right with about 23 million viewers. That skew tells us the typical Office episode drew a pretty consistent audience, and that the big viewership spikes seem to be an exception as opposed to a specific trend.

ggplot(office_ratings, aes(x = imdb_rating)) +
  geom_histogram(binwidth = 0.25) +
  labs(title = "Distribution of IMDb Ratings",
    x = "IMDb Rating",
    y = "Number of Episodes")

IMDb ratings are much more symmetric, peaking around 8-8.5. This makes sense since IMDb ratings mostly reflect how much fans enjoyed the episode itself, which is a more stable measure; unlike live viewership, which can be impacted by other factors like scheduling, promotion, or lead-in programming. There are a few standout episodes rated above 9.3, while the weakest episodes make it into the high 6s, but I wouldn’t say that there are any extreme outliers.

ggplot(office_ratings, aes(x = total_votes)) +
  geom_histogram(binwidth = 250) +
  labs(title = "Distribution of IMDb Votes",
    x = "Number of IMDb Votes",
    y = "Number of Episodes")

The total votes is the most heavily right-skewed of the three variables. Most episodes sit between roughly 1,500 and 2,500 votes, but a few episodes have accumulated 5,000 to nearly 8,000 votes. These are the episodes fans return to and talk about years later, so they seem to keep accumulating ratings long after they aired.

Relationship Between Viewership and Ratings

ggplot(office_ratings, aes(x = viewers, y = imdb_rating)) +
  geom_point() +
  geom_smooth(method = "lm", se = FALSE) +
  labs(title = "Episode Viewership and IMDb Ratings",
    x = "Viewers (millions)",
    y = "IMDb Rating")

There is a positive but only moderate relationship between viewers and IMDb rating. Episodes that drew bigger live audiences do tend to be rated somewhat higher, but the relationship isn’t super strong since there are plenty of modestly watched episodes that earned excellent ratings, and vice versa. Live viewership is shaped by scheduling and promotion as much as by episode quality, so it gives us only a rough idea for how much an episode is actually liked.

Exceptions

highlight_eps <- office_ratings %>%
  filter(title %in% c("Stress Relief", "Pilot"))
ggplot(office_ratings, aes(x = viewers, y = imdb_rating)) +
  geom_point(alpha = 0.4) +
  geom_point(data = highlight_eps, color = "red", size = 3) +
  labs(title = "Notable Exceptions to the Viewership–Rating Trend",
    x = "Viewers (millions)",
    y = "IMDb rating")

Two episodes stand out as clear exceptions. “Stress Relief” drew around 23 million viewers (seemingly because it aired directly after the Super Bowl). Also, the “Pilot” episode, seems to have drawn a large audience but perhaps it was because people were curious about this new show. Both cases show that a big audience can come from outside factors (scheduling, curiosity) rather than from the episode being especially well loved.

Does higher viewership mean more IMDb votes?

ggplot(office_ratings, aes(x = viewers, y = total_votes)) +
  geom_point() +
  geom_smooth(method = "lm", se = FALSE) +
  labs(title = "Viewership and Number of IMDb Votes",
    x = "Viewers (millions)",
    y = "Total IMDb Votes")

Similar to before, the relationship has a positive but moderate correlation, but perhaps it’s a little weaker compared to the previous relationship. In other words, it looks like an episode’s vote count tracks how well-liked and memorable it became over time more closely than it tracks how many people watched it live.

Exceptions

highlight_votes <- office_ratings %>%
  filter(title %in% c("Finale", "Goodbye, Michael", "Stress Relief"))
ggplot(office_ratings, aes(x = viewers, y = total_votes)) +
  geom_point(alpha = 0.4) +
  geom_point(data = highlight_votes, color = "red", size = 3) +
  labs(title = "Notable Exceptions to the Viewership–Vote Count Trend",
    x = "Viewers (millions)",
    y = "Number of IMDb votes")

The series “Finale” is the clearest exception because it drew a relatively normal 5.7 million live viewers, but it has accumulated almost 8,000 IMDb votes, which is clearly the most of any episode, because it’s the episode new fans can watch (and rate) after binge-watching the whole series years later. “Goodbye, Michael” behaves the same way. These episodes’ cultural significance seems to be what drives long-term engagement.

Changes in Popularity Over Time

ggplot(office_ratings, aes(x = air_date, y = viewers)) +
  geom_line() +
  geom_point() +
  geom_smooth(se = FALSE) +
  labs(title = "The Popularity of The Office Over Time",
    x = "Original Air Date",
    y = "Viewers (millions)")

The show initially gained popularity after its first season and generally maintained strong viewership through the middle seasons. Over time, however, original broadcast audiences declined steadily The final seasons had much lower viewership than the peak years. There is also a major spike during season five for Stress Relief, which attracted nearly 23 million viewers (as mentioned earlier).

Changes in Appeal Over Time

ggplot(office_ratings, aes(x = air_date, y = imdb_rating)) +
  geom_line() +
  geom_point() +
  geom_smooth(se = FALSE) +
  labs(title = "The Appeal of The Office Over Time",
    x = "Original Air Date",
    y = "IMDb Rating")

Unlike viewership, the show’s IMDb ratings remained relatively high for much of the series. Although ratings declined somewhat during later seasons, some late episodes received very strong ratings. This demonstrates that popularity and appeal are not the same and that fewer people may have watched the show during its original broadcasts, while the people who did watch could still rate certain episodes very highly.

Comparing Popularity and Appeal

The visualizations show that popularity and appeal did not follow exactly the same pattern throughout the series. Television viewership generally declined over time, especially during the final seasons. However, IMDb ratings remained relatively strong and even increased for certain major episodes near the end of the series. This difference may occur because viewership measures the number of people watching during the original broadcast, whereas IMDb ratings reflect the opinions of viewers who may watch and rate the episodes much later. Therefore, declining broadcast audiences do not necessarily mean that viewers liked the show less.

Viewrship Within Individual Seasons

office_ratings <- office_ratings %>%
  mutate(episode_num = as.numeric(episode))
ggplot(office_ratings,aes(x = episode_num, y = viewers)) +
  geom_line() +
  geom_point() +
  facet_wrap(~ season, scales = "free_x") +
  labs(title = "Episode Viewership Within Each Season",
    x = "Episode Number",
    y = "Viewers (millions)")

Several seasons show some slight changing in viewership within the season. The most dramatic example is Season 5, where Stress Relief had a massive spike in viewers compared with surrounding episodes. The episode aired immediately after the Super Bowl, which helps explain its unusually large audience.

Also, I noticed that the episodes near the end of the series attracted more viewers than nearby episodes. The series finale, for example, experienced higher viewership than most other Season 9 episodes.

Conclusion

Overall, this report analysis shows that the popularity and appeal of The Office changed in different ways throughout the series. The show’s original television viewership generally increased during its earlier seasons before declining gradually during the final seasons. However, IMDb ratings remained relatively high throughout much of the series, highlighting that a decline in broadcast viewership did not necessarily mean that viewers found the episodes less appealing. Several episode exceptions, including Stress Relief, Goodbye, Michael, and the series finale, show how special events and important moments in the series can produce unusually high viewership, ratings, or numbers of IMDb votes. Thus, the data suggests that popularity and appeal should be considered separately when evaluating the overall success of a television series.