Holly Williams, Aakanksha Dimri, and Thomas Pin's DSCI 532 dashboard project on the movies dataset.
Links to important files:
This repo contains the following subfolders:
data: contains all raw data for the dashboardsrc: contains all code files used to wrangle the data, create plots, and make the dashboardimg: contains saved images and plots
Movie studios will often boast about their high box office numbers, but without a larger picture of the entire industry it can be difficult to gage what these numbers really mean. The success of a film can be determined by looking into outcomes such as profits, butts in seats and box office. These numbers are recorded in databases such as IMDB, but this platform doesn’t allow you to compare films for a more holistic picture of the movie industry and what movies are successful.
This visualization app allows users to explore questions such as whether movies are becoming more expensive to make, whether they’re becoming more profitable now than in the past, and whether audiences are going to the movies more or less frequently than in previous years.
The main feature of this app is that it lets users choose which 'success' metric to explore. They can focus on box office sales, profit, butts in seats, or budget and can choose whether to adjust for inflation or not. The app features a large plot (plot 1 in the figure below) showing data for all movies for the period of record (1980 to 2010). The user can choose whether this plot is a line plot (to get a quick snapshot of average change over time), or a box plot (if they are more interested in variability). There are also two bar plots below. The left bar chart (plot 2 in the figure below) has a drop-down menu that lets users explore their own comparisons by selecting movies to compare. The right chart (plot 3) shows the top 10 films and their values for the selected metric.
Overall:
- A drop-down button allows users to select a comparison metric that will be used on all plots. Options include:
- Box office sales
- Profit
- Butts-In-Seats
- A radio button allows users to select between gross values or values that are adjusted for inflation (note: the butts-in-seats metric is measured in number of peple so the plot for this metric will be the same for either selection)
Chart-specific:
- Plot 1 (all movies) has an associated radio button to select whether the data is shown as a box or line chart
- Plot 1 also has a hover tool that allows you to see the movie title, distributer, year, and value for the selected metric
- Plot 3 has a multi-selection drop down menu that will allow users to select as many movies as they like to compare
As mentioned in the proposal, we have excluded:
- movies released before 1980 (due to the presence of erroneous values)
- movies released after 2010 (from lack of information)
- movies in the bottom 5% of
production_budgets - movies in the bottom 1% of
US_Gross
Other limitations related to time constraints include:
- we were unable to remove default plotly hover tool over all plot(s) (even after seeking help from TA's and other colleagues)
- we were unable to apply CSS formatting to correct the colour of the
year slider
Potential future additions / improvements could include:
- add more data from outside resource(s) to compensate for missing years of data
- separate the data into small and big production(s)
- include CSS formatting to improve the overall layout
- add additional groupings to explore patterns amongst distributors or genres of movies (horror, action, romantic comedies, etc)
- improve the code (potentially through vectorization) to reduce delays between making a selection and having the plots update