TheBigScreen

From Visual Analytics for Business Intelligence
Revision as of 17:06, 9 October 2016 by Huiyan.chen.2015 (talk | contribs)
Jump to navigation Jump to search
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

20 Most Profitable Movies.png

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

Duration vs IMDB Score.png

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

Ratings vs Score.png

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
Sunburstsample.png

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
Treemapsample.png

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
Coomments
Linebarscattersample.png

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
|
'

|-

|
Linebarscattersample.png

| 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.

|- |}

Milestone and Schedule

Comments and Feedback