rfordatascience / tidytuesday

Official repo for the #tidytuesday project
Creative Commons Zero v1.0 Universal
6.76k stars 2.39k forks source link

Film Shot Stats/Cinemetrics #363

Open wurli opened 3 years ago

wurli commented 3 years ago

Film Shot Stats/Cinemetrics

The data this week comes from Barry Salt's collection of film cinemetrics. These give information about the average shot length (ASL) of different films, along with many other metrics around camera movement and shot scale.

More information about the data can be found at Barry Salt's own website. Film enthusiasts may also appreciate the use of the data in The Nerdwriter's video on There Will Be Blood.

Credit: Jacob Scott

The data is taken from Barry Salt's website. Cleaning script:

library(dplyr)
library(purrr)
library(readxl)

# http://www.starword.com/Data_Method/Data_Tables/data_tables.html

camera_movement <- map(1:2, ~readxl::read_xls("CamMove.xls", sheet = .)) %>%
  map(rename_with, janitor::make_clean_names) %>%
  map(rename_with, ~"title", any_of("film")) %>%
  map(rename_with, ~"track_w_pan_tilt", any_of("tr_w_pan_tilt")) %>%
  map(rename_with, ~"zoom_w_pan_tilt", any_of("zoom_w_move")) %>%
  bind_rows() %>%
  mutate(across(where(is.double), tidyr::replace_na, 0)) %>%
  distinct()

shot_length <- read_xls("SaltASL5.XLS", sheet = 6, guess_max = 5000, range = cell_cols("A:J")) %>%
  rename_with(janitor::make_clean_names) %>%
  filter(!if_all(everything(), is.na)) %>%
  mutate(
    asl = as.numeric(stringr::str_replace(asl, ",", ".")),
    year = as.integer(year)
  ) 

shot_scale <- map(1:3, ~read_xls("SCALSTAT.XLS", sheet = ., col_types = "text")) %>%
  bind_rows() %>%
  rename_with(janitor::make_clean_names) %>%
  filter(bcu != "BCU") %>%
  mutate(
    year = as.integer(year),
    across(c(asl:vls, -country), as.double)
  ) %>%
  distinct()
wurli commented 3 years ago

Here is a data dictionary compiled from this page.

Camera Movement

variable class description
title character Film title
year integer Film year
director character Director name
pan double Number of panning shots (pans >= 10 degrees)
tilt double Number of tilt shots (tilt >= 10 degrees)
pan_w_tilt double Number of shots with pan and tilt
track double Number of tracking shots
track_w_pan_tilt double Number of tracking shots with pan and tilt
crane double Number of shots using a crane
zoom double Number of zoom shots
zoom w_pan_tilt double Number of zoom shots with pan and tilt
total double Total number of shots

Shot Length

variable class description
title character Film title
alternative_title character Alternative film title
year integer Film year
Director character Director name
asl double Average shot length
country character The name of the country that entirely financed the film if applicable
scope character Scope of the film: 's' means the intended projection aspect ratio of the film is 1:2.2 or 1:2.35 or wider; 't' means that it is a 'made for television' movie; 'v' means that it went straight to video; 'a' means that it is an animated film; 'd' means that it is a documentary.
cinematographer character Cinematographer name
editor character Editor name
studio character The single major studio to finance the film, if applicable

Shot Scale

variable class description
title character Film title
alternative_title character Alternative film title
year integer Film year
director character Director name
asl double Average shot length
ra double Number of reverse angle cuts, which are defined as changes of camera direction of more than 90 degrees
pov double Number of point of view shots, i.e. those taken from the position of one of the participants in the scene, or at any rate from a position on their line of sight
ins double Number of insert shots, i.e. shots of things which do not include an actor's face. Very distant shots of street scenes, in which no principal actors are included, are counted as insert shots
country character The name of the country that entirely financed the film if applicable
intertit double The number of intertitle shots, mainly used in silent films
bcu double Number of Big Close Up shots, i.e. those showing head only
cu double Number of Close Up shots, i.e. those showing head and shoulders
mcu double Number of Medium Close Up shots, i.e. those showing the body from the waist up
ms double Number of Medium Shots, i.e. those showing from just below the hip to above the head of upright actors
mls double Number of Medium Long Shots, i.e. those showing the body from the knee upwards
ls double Number of Long Shots, i.e. those showing at least the full height of the body
vls double Number of Very Long Shots, i.e. those showing the actor small in the frame