The IMDB Data-Set

The data-set can be found on IMDB’s website here.

The format of the data is as follows (from the IMDB website):

Each dataset is contained in a gzipped, tab-separated-values (TSV) formatted file in the UTF-8 character set. The first line in each file contains headers that describe what is in each column. A ‘’ is used to denote that a particular field is missing or null for that title/name.

It took a bit to parse out this data, and drill down to only the films/actors of interest. I may do another post later on displaying the data-pipelining behind this analysis.

James Bond films by lead actor

Whenever a movie series lasts as long as James Bond, different actors play the lead role of the British secret agent. Viewers, immediately go to arguing who was/is the greatest James Bond in the franchise.

Let’s take a look at the ratings submitted to IMDB, on who was the best lead actor to star in a James Bond film.

df %>% 
    group_by(actor) %>% 
    summarise(rating = round(mean(averageRating), 2)
              ,year = min(year)) %>% 
    arrange(desc(rating)) %>% 
    ggplot(aes(x = rating, y = fct_reorder(actor, year), fill = rating, label = round(rating,2))) +
        geom_col() +
        geom_text(nudge_x = .5) + 
        scale_y_discrete(name = NULL, labels = labels) +
        guides(fill = FALSE) + 
        theme(axis.text.y = element_markdown(lineheight = 1.2)) +
        labs( title = 'Bond Films'
             ,subtitle = 'Average Rating by Actor'
             ,caption = 'Source: IMDB.com'
             ,x = 'Average Rating')

According to the ratings found on IMDB, the highest rated actor is Daniel Craig, with an average rating of 7.28, followed closely behind by Sean Connery at 7.13. The worst actors (according to the average rating) were Pierce Brosnan at 6.55, and Timothy Dalton at 6.65.

Top Rated Films

It is no surprise, when looking at the top rated leads, that films that Daniel Craig has starred in rank among the top rated films in the series. Casino Royale and Spectre (both films Craig starred in) come in at number 1 and 3 respectively, with Sean Connery accounting for the other 3 spots.

df %>% 
    group_by(title = originalTitle) %>% 
    summarise(rating = round(mean(averageRating), 2)
             ,year = min(year)
             ,budget = mean(budget_adjusted)
              ) %>%
    mutate(title = fct_reorder(title, year)) %>% 
    ggplot(aes(x = rating, y = title, fill = rating, label = rating)) + 
    geom_col() + geom_text(nudge_x = .25) + guides(fill = FALSE) +
    labs( x = 'Average Rating'
         ,y = ''
         ,title = 'Average Rating of Bond Films'
         ,subtitle = 'Ordered by Release Year')

df %>% 
    arrange(desc(averageRating)) %>% 
    filter(row_number() < 6) %>% 
    group_by( Title = originalTitle
             ,`Lead Actor` = actor) %>% 
    summarise(`Average Rating` = round(mean(averageRating), 2)
             ,Year = min(year)
              ) %>%
    arrange(desc(`Average Rating`)) %>% 
    kableExtra::kable(caption = 'Top 5 rated James Bond films')

Budget

Bond films are notorious for their spectacular stunts, and massive explosions. Not surprisingly, this also means huge budgets for the films. As the years have gone on, the budgets have increased dramatically.

The graph below displays the budget for each film, plotted against the year it was released, with the average rating of the film indicated by the color.

I have adjusted for inflation, by scaling the budgets to equivalent 2005 USD.

fit.1 <- lm(budget_adjusted ~ year, data = df)

df %>% 
    group_by(title = originalTitle) %>% 
    summarise(rating = round(mean(averageRating), 2)
             ,year = min(year)
             ,budget = mean(budget_adjusted)
              ) %>% 
    ggplot(aes(x = year, y = budget
             , size = rating, color = rating, label = title)) +
        geom_abline( intercept = coef(fit.1)[1]
                    ,slope = coef(fit.1)[2]
                    ,color = 'Red'
                    ,size = 1) +
        geom_point() +
        geom_text_repel(aes(x=year,y=budget,label=title)
                , inherit.aes = FALSE
                ) +
        guides(size=FALSE) +
        labs( y = 'Budget (USD Millions)'
             ,x = 'Year'
             ,color = 'Rating')

There is a clear positive relationship between release year, and budget, with the release year accounting for over 86% of the variance in budget. On average, for every year that goes by, the budget for a James Bond film increases $3.335 Million dollars!

Interesting to note, however, that there is no relationship between budget, and average rating for a film.

fit.2 <- lm(budget ~ averageRating, data = df)

df %>% 
    group_by(title = originalTitle) %>% 
    summarise(rating = round(mean(averageRating), 2)
             ,year = min(year)
             ,budget = mean(budget_adjusted)
              ) %>% 
    ggplot(aes(x = rating, y = budget
             , size = rating
             , color = rating
             , label = title)) +
        geom_abline( intercept = coef(fit.2)[1]
                    ,slope = coef(fit.2)[2]
                    ,color = 'Red'
                    ,size = 1) +
        geom_point() +
        geom_text_repel(aes(x=rating,y=budget,label=title)
                , inherit.aes = FALSE
                ) +
        guides(size=FALSE) +
        labs( y = 'Budget (USD Millions)'
             ,x = 'Average Rating'
             ,color = 'Rating')

Conclusion

There is a ton of information in the IMDB dataset, that is a great start for any data science project for the movie or television lover. There are still many other questions we could investigate with this data, and more posts may be coming about the James Bond series.