Week 12 Bechdel

Your RMarkdown notebook for this data dive should contain the following:

• Select a column of your data that encodes time (e.g., “date”, “time stamp”, “year”, etc.). Convert this into a Date in R. - Released

  bechdel_data_movies$released_as_date <-  as.Date(bechdel_data_movies$released, format = "%d %b %Y")
  view(bechdel_data_movies$released_as_date)

• Choose a column of data to analyze over time. This should be a “response-like” variable that is of particular interest. - binary

Create a tsibble object of just the date and response variable. Then, plot your data over time. Consider different windows of time.

bechdel_ts <- bechdel_data_movies |>
  filter(!is.na(released_as_date)) |>
  select(released_as_date, budget, imdb_id) |>
  as_tsibble(key = imdb_id, index = released_as_date)

bechdel_ts

## # A tsibble: 1,591 x 3 [1D]
## # Key:       imdb_id [1,591]
##    released_as_date   budget imdb_id
##    <date>              <dbl> <chr>  
##  1 1970-06-17        1000000 0065466
##  2 1971-05-21        2500000 0067065
##  3 1971-10-09        2200000 0067116
##  4 1971-07-02       53012938 0067741
##  5 1971-12-29       25000000 0067800
##  6 1972-11-17        4000000 0068156
##  7 1972-03-24        7000000 0068646
##  8 1973-08-22       15700000 0068699
##  9 1979-10-01          12000 0069089
## 10 1973-08-11         777000 0069704
## # ℹ 1,581 more rows

budate_graph <- ggplot(bechdel_ts, aes(x = released_as_date, y = budget)) + geom_point() + geom_smooth(method = 'lm', color = 'red', se = FALSE) + labs (title = 'Movie Gross Unadjusted Budget by Release Date')
budate_graph

## `geom_smooth()` using formula = 'y ~ x'

• What stands out immediately?

  • Budgets appear to increase over time, which is expected given inflation, but it appears to outpace inflation.

• Use linear regression to detect any upwards or downwards trends.

• The positive trend of roughly $1.3 million per year ($3,569.83 per day) likely reflects a combination of inflation, rising production standards, expensive technological advances, and the industry’s increasing emphasis on high‑budget films.

  budget_model <- lm(budget ~ released_as_date, data = bechdel_ts)
  
  budget_model$coefficients

  ##      (Intercept) released_as_date 
  ##      1940050.057         3569.834

  budget_model2 <- loess(budget ~ as.numeric(released_as_date), data = bechdel_ts, span = .5, na.rm = TRUE)
  
  budget_model2

  ## Call:
  ## loess(formula = budget ~ as.numeric(released_as_date), data = bechdel_ts, 
  ##     span = 0.5, na.rm = TRUE)
  ## 
  ## Number of Observations: 1591 
  ## Equivalent Number of Parameters: 6.99 
  ## Residual Standard Error: 46640000

  • There is a modest upward trend.

    • Do you need to subset the data for multiple trends?

      • No

    • How strong are these trends?

      • Fairly strong; the increase in budgets over time is large enough that it’s unlikely to be random noise.

• Use smoothing to detect at least one season in your data, and interpret your results.

• budate_graph2 <- ggplot(bechdel_ts, aes(x = released_as_date, y = budget)) + geom_point() + geom_smooth(method = 'loess', span = .75, se = FALSE) + labs (title = 'Movie Gross Unadjusted Budget by Release Date with Loess')
  budate_graph2

  ## `geom_smooth()` using formula = 'y ~ x'

  

  • There doesn’t appear to be seasonality in this data, even when fitted with a loess curve with a high span. I expected to see some wiggle within years to account for things like summer blockbusters and holiday movies, but that does not appear to be the case. There is a steady rise until about 2008, where the rate of increase gets somewhat steeper.

  • Can you illustrate the seasonality using ACF or PACF?

    • No, there appears to be no seasonality. This is confirmed by the ACF, which has all near-zero values after one spike at 0. The PACF also suggests there is no auto regression, and thus no seasonality, additionally, the spikes are irregularly placed. This means there is no statistically significant seasonality in movie budgets in the data set.

acfbechdel <- acf(bechdel_ts$budget)

acfbechdel

## 
## Autocorrelations of series 'bechdel_ts$budget', by lag
## 
##     0     1     2     3     4     5     6     7     8     9    10    11    12 
## 1.000 0.085 0.054 0.043 0.086 0.062 0.057 0.040 0.072 0.096 0.036 0.050 0.031 
##    13    14    15    16    17    18    19    20    21    22    23    24    25 
## 0.031 0.048 0.100 0.043 0.059 0.041 0.068 0.043 0.057 0.063 0.073 0.090 0.053 
##    26    27    28    29    30    31    32 
## 0.074 0.047 0.055 0.058 0.050 0.027 0.063

pacfbechdel <- pacf(bechdel_ts$budget)

pacfbechdel

## 
## Partial autocorrelations of series 'bechdel_ts$budget', by lag
## 
##      1      2      3      4      5      6      7      8      9     10     11 
##  0.085  0.047  0.035  0.078  0.046  0.040  0.023  0.055  0.075  0.009  0.029 
##     12     13     14     15     16     17     18     19     20     21     22 
##  0.006  0.005  0.027  0.078  0.014  0.033  0.012  0.039  0.012  0.030  0.037 
##     23     24     25     26     27     28     29     30     31     32 
##  0.037  0.054  0.017  0.039  0.012  0.018  0.024  0.008 -0.010  0.026