Team Shooting Stars: Proposal

From Visual Analytics for Business Intelligence
Revision as of 19:54, 2 December 2016 by Manasm.2012 (talk | contribs)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search
Group4 teamlogo.jpg




Background & Motivation

The National Basketball Association (NBA) is one of the most famous sports league in the world. It consists 30 men basketball teams (where 29 in the US, 1 in Canada) which was founded in 1940s, named BAA (Basketball Association of America). Then it changed to the current name of NBA after merging with NBL (National Basketball League) in 1949. NBA plays are generally fast-paced, physically intensive where audience find it fascinating to watch. NBA also represents the best basketball play standard in the world. Joining NBA is the ultimate dream for a professional basketball player.

As NBA fans, our group would like to analyse on player’s statistics and team’s performance to clear our doubts like how the play styles of NBA basketball have been changed over the last 10 years. We would apply different visualization tools and graphics to gain in-depth analysis.

Project Description

The aim of this project is to get deeper knowledge into the current trend in NBA and what makes a team succeed, so the questions we are going to answer are:
1. Is the role of centre becoming less and less important in NBA?
2. How do the defense and offense factors of a team vary and determines a team's success?
3. What are the important quality that leads to a player's success?
4. Does the presence of a superstar and a bigger budget lead to the success of the team?


Data Set Selection

We retrieved our data from Basketball Reference. The data is in CSV format where each game contains two CSVs files. For example, the following two CSVs represent the box scores of Cleveland Cavaliers vs Golden States Warriors on June 19, 2016:

Raw data 01.jpg


Raw data 02.jpg


Moreover, we also can retrieve a specific player's game statistics in a certain timeline from this site:

Raw data 03.jpg


For our VA project, we plan to retrieve all players data in the past 10 years. We would also categorize the game statistics according to the game type (normal season, playoff, finals). The data size is quite large so we will use JMP to do data transformation and combination.

Schedule

Academic Studying Week Task Done By Status
Week 7
1 Brainstorm project topic and scope Everyone Completed
Week 8
1 Formulate ideas Everyone Completed
2 Consulting Prof Wu Wei, Wang Ziteng Completed
3 Deciding on tools/techniques to use Wang Ziteng, Manas Mohapatra Completed
4 Upload detail project proposal Wu Wei Completed



Week 9
1 Data preparation, consolidation, preprocessing and cleaning Everyone Completed
Week 10
1 Update Wiki page Wu Wei, Wang Ziteng Completed
2 Study Treemap Wang Ziteng Completed
3 Study Multi-Series Line Chart Manas Mohapatra Completed


Week 11 & 12
1 Web App Developer Everyone Completed
Week 13
1 Do poster, presentation preparation Everyone Completed
Week 14
1 Presentation Everyone Completed
Week 15
1 Submission of project poster Wu Wei Completed
2 Submission of final project paper and artifacts Wang Ziteng, Manas Mohapatra Completed
Week 16
1 Visual Analytics Poster Night Everyone Completed

Tools

Our team decided to use the tools such as JMP, Tableau, d3,js for doing the following analysis:

Illustration Analytical Methods
TSS2.png


  • Spider Charts
  • Logistic Regression: The outcome of the analysis will usually be based on a success or failure of the game. Hence to do a regression analysis on factors for success of a game, we need to use logistic regression as the dependent variable is categorical in nature
TSS3.png
  • Time Series Analysis: Basketball Reference consists of 5 years’ worth of dataset. Thus it’s worth doing a time series analysis on player performance, team performance, change in player’s role and several other factors
TSS1.png
  • Treemap
  • Sunburst Charts
  • Heatmaps: Heatmaps are a two-dimensional representation of data in which values are represented by colors. It can provide an instantaneous glance of the summary of information. Given the hyper-dimensional data sets of NBA players and teams, Heatmaps is a effective tool to visualise the dataset

Prior Work & References

  1. Success Factors in NBA: http://www.basketball-reference.com/about/factors.html
  2. The length and success of NBA careers: Coates, Dennis; Oguntimein, Babatunde. International Journal of Sport Finance 5.1 (Feb 2010): 4-26.
  3. Racial Discrimination among NBA referees: http://ww2.amstat.org//Chapters/boston/nessis07/presentation_material/Justin_Wolfers.pdf