TheBigScreen

Proposal

Poster

Application

Research Paper

Problem and Motivation

Is it possible to predict how good a movie will be before it even screens? This is a subjective question. While some rely on movie critics and early reviews, others depend on instinct. However, we know reviews can take a long time to gather and human instinct is simply unreliable. Thousands of movies are produced every year and all of them our clamouring for the $11 we spend on movie tickets! Our group wants to know if we can predict which movies are worth you spending your money and time on.

Data

We are using the IMDB 5000 Movie Dataset from Kaggle. The Internet Movie Database (IMDB) is an online database of information related to films, television programs and video games [1]. Amongst its functions, IMDB allows users rate movies on a scale of 1 to 10.

The dataset contains the following variables, including but not limited to:

• movie title
• director name
• actors’ names and Facebook likes
• length of movie
• year
• gross earnings
• genres
• language
• country
• content rating
• budget
• IMDB rating

Related Work

Visualizations	Learning Points
Top 20 Most Profitable Movies Source: https://www.kaggle.com/param1/d/deepmatrix/imdb-5000-movie-dataset/the-money-makers	Simple and easy to read Not aesthetically pleasing Data points not properly explained e.g. why do some points have tails
Duration of Movie vs. IMDB Score Source: https://www.kaggle.com/benjaminlott/d/deepmatrix/imdb-5000-movie-dataset/imdb-5000-general-data-analysis	Colours are visually appealing Messy Visualization was so big that legend could not fit on the same window Tooltip tags are unformatted and messy
Age Ratings vs IMDB Score Source: https://www.kaggle.com/adhok93/d/deepmatrix/imdb-5000-movie-dataset/eda-with-plotly	Appropriate and informative use of boxplots to visualize continuous variable, IMDB scores Messy, especially when tooltip is displayed Unnecessary legend and use of colours No y-axis title

Technical Challenges

Challenges	Approach
Data Cleaning & Exploration	Collaborate in data cleaning and transformations
Use of Javascript and D3	Attend coding workshop in recess week Consult with Prakash
Implementing Interactive Visualizations	All team members to explore the use of various visualization tools e.g. Tableau, Power BI, Qlik Sense Explore how to implement interactivity and animation through online tutorials

Method of Approach

Sunburst Chart	Comments
	Features: For a breakdown of Genres Combination of Genres Actors and Actresses Director Size of each portion illustrates figures for the category such as grossing, number of likes on social media platform. Our team felt that a Sunburst chart will be useful to visualize the data as users will be able to go deeper and understand the the data as they break it down further. It is important to note that a detailed guide for its usage is needed as this is a rather uncommon visualization chart.
TreeMap	Comments
	Features: Able to identify weights of each category by size and color (highest grossing genres, largest area) Easily understood by general public, using only simple sizing and colors to differentiate characteristics of the categories. Size of each portion illustrates figures for the category such as grossing, number of likes on social media platform. Our team felt that a Treemap will be useful for delivery insights of the data easily to everyone. As compared to visualizations like sunburst where users are required to learn about interacting with the chart for more insights, Treemaps deliver them effortlessly and usage is also simple.
Line/Bar/Scatter	Comments
	Features: Offers great visualization of relationship between variables Displays much clearer figures for each variable, interactiveness is highly customizable Commonly used therefore generally easy to understand for readers The commonly used and still popular line/bar/scatter charts displays visualization that allows readers to accurately determine relationship between categories. These charts will be frequently used by us to present figures such as grossing against budget, movie success and social media popularity.

Proposed Storyboard

First Proposed Storyboard

'

Our first proposed storyboard includes line/bar/scatter chart that works together like the Tablaeu dashboard, the Sunburst diagram and the Treemap. This initial mockup will be done in tableau allowing us to easily try out different combinations of variables and see the kind of results we will get. The line/bar/scatter chart will be used for showing if a relationship exist between variables and if they are correlated strongly. We plan to experiment on the following but not limited to: Grossing vs Budget Social media popularity or actors vs Movie ratings Genre vs popularity/grossing/budget The Sunburst diagram will let users breakdown deeper into each category based on: Genre (possibly multiple combinations of genre) Directors Grossing Viewership Ratings/Critic scores The Treemap will

Milestone and Schedule

Comments