MAT022 Foundations of Statistics and Data Science Summative Assessment 2020/21
MAT022 Foundations of Statistics and Data Science Summative Assessment 2020/21
Summative assessment for the module is by means of a single report on your statistical anal- ysis of data related to the National Basketball Association (NBA), a men’s professional basketball league in the USA.
This form of assessment has been chosen because as professional statisticians and data scientists, you will often be asked to investigate a data set and report on whether it contains anything useful or interesting. The assessment will also help you to prepare for writing your MSc dissertation in the summer.
Assessment type Weight Max. length Format
Report 100% 10 pages R Markdown and PDF Your report will be assessed according to how well you are able to
• analyse the data set,
• interpret the results of your analysis, and
• present the results of your analysis and your interpretation of the data set. 30%
Your analysis should be performed using the R statistical software package, and your report prepared using the R Markdown typesetting system and the template provided.
1 The data
You are asked to write a report on data from the 2014–15 season of the National Basketball Association (NBA), a men’s professional basketball league in the USA. The data set is a record of all shots taken by players in the NBA between October 2014 and March 2015, and consists of 128,069 observations on 23 variables as described in Table 1.
Deadline
Tue 09 Feb 2021
40% 30%
Please submit your report in PDF format, together with the R Markdown file used to generate the report, through Learning Central sometime before 12.00 on Tuesday 09 February 2021.
1
MAT022 Foundations of Statistics and Data Science Summative Assessment 2020/21
Variable Description
GAME_ID
DATE
HOME_TEAM
AWAY_TEAM
PLAYER_NAME
PLAYER_ID
LOCATION
W
FINAL_MARGIN
SHOT_NUMBER
PERIOD
GAME_CLOCK
SHOT_CLOCK
DRIBBLES
TOUCH_TIME
SHOT_DIST
PTS_TYPE
SHOT RESULT
CLOSEST_DEFENDER
CLOSEST_DEFENDER_ID
CLOSE_DEF_DIST
FGM
PTS
Unique id number of the game.
Date of the game.
Team playing at home.
Team playing away from home.
Name of the shooting player.
Unique id number of the shooting player.
Whether the player was on the home (H) or away (A) team. Whether the player’s team won (W) or lost (L) the game.
The margin of victory for the player’s team (negative means defeat). The number of the shot taken by the shooting player in that game. The period of the game that the shot was taken.
The time remaining in the period when the shot was taken. The time remaining on the shot clock.
Number of dribbles by the player before the shot was taken. The time that the ball was in the shooting player’s hand. The distance of the shooting player from the basket.
2 for shots from inside the arc, 3 for shots from outside the arc. Whether the shot was successful (‘made’) or unsuccessful (‘missed’) Name of the nearest defender when the shot was taken.
Unique id number of the nearest defender.
Distance of the nearest defender when the shot was taken. Equal to 1 if the shot was made (scored) otherwise 0.
The number of points scored (0, 2 or 3)
Table 1: Description of the variables in the NBA shot logs data set.
ATL Atlanta Hawks BKN Brooklyn Nets BOS Boston Celts
CHA Charlotte Hornets CHI Chicago Bulls
CLE Cleveland Cavaliers DAL Dallas Mavericks DEN Denver Nuggets DET Detroit Pistons
GSW Golden State Warriors HOU Houston Rockets
IND Indiana Pacers
LAC Los Angeles Clippers LAL Los Angeles Lakers MEM Memphis Grizzlies
MIA Miami Heat
MIL Milwaukee Bucks
MIN Minnesota Timberwolves NOP New Orleans Pelicans NYK New York Knicks
OKC Oklahoma City Thunder ORL Orlando Magic
PHI Philadelphia 76ers
PHX Phoenix Suns
POR Portland Trail Blazers SAC Sacramento Kings
SAS San Antonio Spurs
TOR Toronto Raptors
UTA Utah Jazz
WAS Washington Wizards
Table 2: Acronyms for the teams in the NBA.
2
MAT022 Foundations of Statistics and Data Science Summative Assessment 2020/21
2 The report
The ability to write clearly and concisely is an important professional competence. To encourage writing that is brief and to the point, your reports are limited to a maximum of 10 pages. It is often far more difficult to express yourself in 100 words than in 1000 words, especially when you have a lot to say, so be careful not to underestimate the challenge posed by this restriction. The modest page limit will also encourage you to be selective in the results you choose to present.
A suggested structure for your report is shown in Table 3. Note that the title page, abstract, table of contents, list of references and appendix do not contribute towards the page count.
Title
Abstract
Table of contents
1 page 100 words –
1. Introduction
2. Background
3. (descriptive analysis) 4. (inferential analysis) 5. (inferential analysis) 6. (inferential analysis) 7. Conclusion
1/2 page 1 page 2 pages 2 – 3 pages 2 – 3 pages 2 – 3 pages 1/2 page
References – Appendices 2 pages max.
Table 3: Suggested report structure
• The title page should contain the title of your report, your name and student number, and the date on which your report was completed.
• The abstract should contain a short summary of the report and its main conclusions.
• The table of contents should list the number and title of each section against the number
of the page on which the section begins.
• The introduction should consist of a few short paragraphs, describing the purpose of the report and providing a brief outline of its contents.
• The background section should include a brief review of any relevant literature, and provide a context for the work presented in the report.
• The report should contain a relatively short section on a descriptive analysis of the data set, with a title chosen to reflect what the section contains.
• The main part of the report should consist of two or three sections on different inferential analyses of the data set. Here you should formulate hypotheses, conduct statistical tests, then present and discuss the results of these tests. The titles of these sections should reflect what the sections contain.
• The conclusion should consist of a few short paragraphs, providing a summary of the report and a brief outline of some ideas for future work.
• The report may contain a single appendix for large figures and tables, limited to a maximum of two pages.
3
MAT022 Foundations of Statistics and Data Science Summative Assessment 2020/21
3 Assessment criteria
Detailed assessment criteria are shown in Table 4.
Level
Analysis
(40%)
Discussion
(30%)
Presentation
(30%)
Distinction
(70–100)
Hypotheses are inter- esting and original. Methods are appro- priate and applied carefully and precisely. An interesting de- scriptive analysis is included and reported correctly.
Inferences are valid and supported by evidence. Original and interesting conclusions are articulated. There is some shrewd spec- ulation about possible causal factors.
A high standard of writing is maintained throughout. The nar- rative is clear, coher- ent, eloquent and re- fined. Figures and tables are used cre- atively.
Merit
(60–69)
Hypotheses are formu- lated correctly. Meth- ods are appropriate and applied correctly. A moderately interest- ing descriptive analy- sis is included and re- ported correctly.
Inferences are valid and supported by evidence. Interesting conclusions are artic- ulated. There is some speculation about possible causal factors.
A good standard or writing is maintained throughout. The nar- rative is clear and co- herent. Figures and ta- bles are used to illus- trate the narrative.
Pass
(50–59)
Hypotheses are for- mulated correctly. Methods are applied correctly for the most part. A descriptive analysis is included and reported correctly.
Inferences are mostly valid and supported by some evidence. Some relatively inter- esting conclusions are articulated.
An acceptable stan- dard of writing is maintained through- out. The narrative is lacklusture and sometimes unclear. Figures and tables do not always illustrate the narrative.
Fail
(0–49)
The analysis is bland and almost entirely de- scriptive.
Inferences are invalid or not supported by ev- idence. There is little of any interest.
The report is poorly written. The narrative is disjointed and hard to follow.
Table 4: Assessment criteria
Plagiarism
You may find existing studies of the NBA data set online. Plagiarism is to present other people’s work or ideas as your own, by incorporating it into your work without full acknowledgement. The need to acknowledge others’ work applies not only to text, but also to computer code, figures, tables etc. You must also attribute text, data, or other resources downloaded from websites. Following submission your report will be analysed by the TurnitIn software, and any report in which plagiarism is detected will receive a mark of zero.
4
MAT022 Foundations of Statistics and Data Science Summative Assessment 2020/21
4 Guidelines for writing reports
The golden rule when writing is to always think of the reader. For scientific reports, readers will typically want to read something interesting and learn something in the process.
Not interesting Quite interesting
Very interesting
What do we mean by interesting?
The average exam mark of statistics and data science students.
The average mark of male students, the average mark of female students, and the results of a test of whether any difference is statistically significant. The average mark of male students, the average mark of female students, a statistical test of whether any difference is significant, and some speculation about why there is a significant difference, or alternatively why there is not.
Audience. The target audience for your report is this year’s cohort students on the Founda- tions of Statistics and Data Science module, so you can assume that your readers are familiar with the methods and terminology established within the lectures and notebooks. If you choose to use methods that have not been covered in lectures, you must ensure that any new terms are properly defined and references to the relevant literature included.
Analysis. The reader shoud be satisfied that you have performed your analysis correctly, and in particular that you have verified the conditions that are necessary to apply the various methods. Your methods should be introduced with a brief summary of their main features, but technical details should not be discussed at length although you might consider providing the interested reader with references to the relevant literature.
Navigation. Do not assume that the reader will read the report from start to finish, as one might read a novel. Reports should be made easy to navigate using numbered sections and subsections together with cross-referencing. Once you have written a first draft, it will need careful editing before it becomes a coherent and polished report. This stage always takes longer than you think!
Scientific writing. For scientific reports we aim for a style of writing that is clear and concise. Make sure that sentences are unambiguous and that a good standard of writing is maintained throughout the report.
• Sections should not start abruptly with the subject matter, but rather with an introductory sentence or short paragraph. Sections should also end with concluding sentence or short paragraph.
• All figures and tables must be numbered and have captions. Figures or tables that are not mentioned at least once in the text should not be included.
• A qualified statement is one that express some level of uncertainty about its own accuracy, and should always be used when drawing conclusions from the results of a statistical analysis, and especially when speculating about possible causal factors. Common phrases that indicate qualified statements include “This suggests that …”, “It appears that …”, “We might conclude that …”, “There is some evidence to indicate …” and so on.
5