CS145: Data Management and Data Systems

Stanford University, Fall 2019

Project 1: ​Exploring NCAA Basketball Data

Due Date: Friday, October 11th, 11:59PM

Overview
Welcome to CS145! Throughout the course you will be using Google’s ​BigQuery platform to
gain hands-on practice with real-world data systems.

BigQuery is Google’s service for big data. Throughout the three class projects, we will be using
BigQuery’s basic SQL querying interface, its interface with Colaboratory1, and its built-in
machine learning features.

Google has published many datasets on BigQuery -- these range from StackOverflow statistics to
real-time air quality data. In this first part of the course project you will be using BigQuery’s
SQL interface to answer questions about the NCAA Basketball Dataset. To find the dataset in
BigQuery, follow the instructions in the ​Getting Started with BigQuery​ support document.

Please note: ​this is a solo project​. You may discuss ideas at a high-level with other students, but
all work should be your own. Please note the names and SUNet IDs of any students you
collaborate with.

Task A: Getting Set Up

Note: ​This is a ​very important step​ as we will be unable to give extra Google Cloud credit to
students that use up all of their credit. If you have any questions about Google Cloud, your
account, or your credits, check Piazza for similar questions or make a post yourself.

1
Colaboratory is a free document collaboration tool built on top of ​Jupyter​. You can think of it as a Jupyter
notebook, except that you are able to collaborate on it with multiple people.
1
Task B: Familiarize yourself with the NCAA Basketball Dataset
Now that you’ve oriented yourself in BigQuery, your second task is to examine the schemas and
the descriptions of the NCAA Basketball dataset tables and understand the data that you will be
working with.

You may try running some simple queries over the tables to get a feel for them, or use
BigQuery’s “Preview” tab to see what the data looks like.

Some notes:
● _sr stands for “Sportradar”, which is a company that collects sports data, down to the x/y
coordinates of events (shot attempted, rebound, turnover, foul).
● The historical data makes a distinction between tournament games and regular season
games. Please make sure you're using the right table!
● “pbp” means play-by-play, which is very granular data about each event that happens in
the game

Task C: Querying!
Now that you’ve gotten comfortable and familiar with BigQuery and its SQL querying interface,
let’s get to work and answer some questions about the NCAA Basketball dataset.

We intend for part of this assignment to be about how to translate a question in plain English to a
schema - in other words, we want you to read the tables and explore the data and think about
which tables and columns are necessary in answering the question we’re asking. This skill is
both necessary for the remainder of the projects, and is exactly how real world data querying and
analysis works!

2
Your queries should be fairly efficient -- they should each take at most ten seconds to execute on
BigQuery, and most of them will be finished in less than ~4 seconds. If any of your queries are
taking much longer than that, you’ve probably written them in particularly inefficient way;
please try rewriting them, and see the course staff if you need help.

Please also check to make sure you’re not querying more than a couple GBs of data - we’ve
specifically chosen this dataset so that no one need exhaust their credits completing the
assignment. All the queries you write should fit within the 1TB of free querying you’re alloted
for the month.

You can save your queries for each question from the BigQuery interface directly, or you can
keep track of your queries in separate files yourself. Remember that you can use BigQuery’s
“Query History” tab to inspect previous queries you’ve run.

Note: ​When querying in BigQuery, ​table names should be wrapped in backticks (`). For
example, instead of saying:
SELECT * FROM bigquery-public-data.ncaa_basketball.mascots
say:
SELECT * FROM `bigquery-public-data.ncaa_basketball.mascots`

Questions:

We will provide answers for these questions so that you can check your work. ​Please make sure
your output from BigQuery matches these answers​, both in terms of values and ordering.
Read instructions carefully; if we ask for rounded answers, we may deduct points for not
rounding.

Note: ​While matching these answers is a good sanity check, it does not guarantee a perfect
score. The datasets we will use to grade your assignment may not perfectly match the datasets
on BigQuery; therefore, make sure that your queries are generalizable to other datasets (given
that schemas are identical).

We reserve the right to deduct points from your project if your queries are hard-coded in some
way or are not generalizable to other tables.

For the following questions, unless otherwise specified, a game can be either a tournament
game or a regular season game.

Write standard SQL queries to answer the following questions:

1. (1 point) What is the name and capacity of Stanford’s NCAA basketball team venue?
3
 Answer:
Row venue_name venue_capacity
1 Maples Pavilion 7392

2. (1 point) How many games were played in Stanford’s venue in the 2013-2014 season?

Answer:
Row games_at_stanford
1 16

3. (1 point) Hexadecimal colors codes are a way of representing color on a computer. Hex
color codes are of form #AABBCC, where AA, BB, and CC are hexadecimal numbers
(00, 01, … , FE, FF) indicating the intensity of red, green, and blue in the color,
respectively.

Hint​: be careful with the case of the colors in the dataset -- some use lower case
characters and some use upper case characters. Note that in the expected answer below,
the original case from the dataset is kept.

What teams have the maximum possible red intensity in their color? Give (team market,
color) as your answer. Order your results alphabetically by the team name.

Answer:
Row market color
1 Idaho State #ff7800
2 Morehead State #ffc300
3 North Carolina A&T #ffb82b
4 Northern Colorado #ffb500
5 Oklahoma State #FF6600
6 Pacific #ff6900
7 South Dakota #ff2310
8 Syracuse #ff5113
9 Tennessee-Martin #ff6900

4. (1 point) How many ​home ​games has Stanford won in seasons 2013 to 2017 (inclusive)?
Give (number of games won, average score for Stanford in those games, average score of
the opponents in those games) as your answer. Round any decimal values to two places.

4
 Answer:
Depending on which table you use for your query, you may get slightly different values.
Either of the following results are acceptable.

Row number avg_stanford avg_opponent

Row number avg_stanford avg_opponent

5. (2 points) How many players have been on a team based in the same city where they
were born? For this question, please only use the player’s birth city and state (do not
include the player’s birth country).

Answer:
Row num_players
1 606

6. (2 points) What is the biggest margin of victory in the historical tournament data? Output
the winning team name, losing team name, winning team points, losing team points, and
the win margin of that game.

Answer:
Row win_name lose_name win_pts lose_pts margin
1 Jayhawks Panthers 110 52 58

7. (3 points) In a basketball tournament, teams are ranked from best to worst prior to
starting the matches. This ranking is called the “seed” of the team (1 is the best team, and
a higher number indicates a worse team). In general, a higher ranked team is expected to
beat a lower ranked team.

Definition​: An ​upset ​occurs whenever a team with seed ​A ​beats a team with seed ​B,​ and
A > B.

5
 What percentage of historical tournament games are upsets? Round to two decimal
places. For example, if 50.2489% of games are upsets, your query should return 50.25.

Answer:
Row upset_percentage
1 27.26

8. (3 points) Which pairs of NCAA basketball teams are 1) based in the same state and 2)
have the same team color? Output the team names and the state. Put the team name that
comes alphabetically first in each pair on the leftmost column, and order the rows
alphabetically by the first column.

Answer:
Row teamA teamB state
1 Bearcats Norse KY

2 Cougars Red Raiders TX

3 Razorbacks Red Wolves AR

9. Definition​: A ​geographical location​ ​L ​is a unique tuple (city, state, country).

Definition​: A geographical location ​L ​“makes” points for a team ​T ​whenever a player that
was born in ​L s​ cores points for ​T.

(3 points) What three geographical locations made the most points for Stanford’s team in
seasons 2013 through 2017, and how many points did they make?

Restrictions:
- For the purposes of this query, avoid using the “​birth_place”​ column.

Answer:
Row city state country total_points
1 Phoenix AZ USA 2223

2 Minneapolis MN USA 1427

3 Rock Island IL USA 1399

6
10. (4 points) Since the start of the 2013 season, which teams have had ​more than 5 players
score 15 or more points in the first half in a single game? ​Note​: These players did not all
have to score 15+ points in the first half of the ​same ​game.

Output the top 5 team markets and the number of players for each team meeting this
criteria from most to least, breaking ties by team markets in alphabetical order.

Answer:
Row team_market num_players
1 Kentucky 14

2 Oregon 14

3 UCLA 14

4 Duke 13

5 Marquette 13

11. Definition​: Team ​X i​ s a ​top performer ​on season ​Y ​if no o​ ther team had more wins than
X ​in the same season. This includes teams with either null or non-null markets.

(4 points) What ​five teams (identify them here by their “markets”) were top performers in
the most seasons between 1900 and 2000 (inclusive), and how many times were they top
performers? Output the team markets and the number of times each team was a top
performer. If there are ties in the final output, break them by giving a higher ranking to
team markets that come first alphabetically. Ignore teams with NULL markets only in the
final output.

Answer:
Row team_market top_performer_count
1 University of California, Los 6
Angeles

2 University of Kentucky 6

3 Texas Southern University 5

4 University of Pennsylvania 5

5 Western Kentucky University 5

7
Submission Instructions
Once you have written queries that answer all questions and conform to the given result schemas,
you’re ready to submit.

To submit:
1. Copy all the queries you wrote for Task C into the ​project1_submission.py ​file
(available on the course website), pasting all of your queries into the corresponding
places.
2. If you collaborated with others to generate your queries, add their names and SUNet IDs
to the comment at the top of the ​project1_submission.py​ file.
3. Submit this Python file on Gradescope. ​In order to be correctly graded, the file must
be named ​project1_submission.py​.

You may resubmit as many times as you like; however, only the latest submission and timestamp
will be saved, and we will use your latest submission for grading your work and determining any
late penalties that may apply. Submissions via email will not be accepted!

IMPORTANT SUBMISSION NOTES:

When you submit to Gradescope, we will run a syntax checker that will make sure that your
SQL runs OK. It should run immediately and return whether the query ran OK or if there were
errors - please ​make sure that you get a positive result from this test in your final
submission.

You will not see a final grade until after the project deadline.​ The answers are provided
above for the questions so that you may check your work yourself. It is your responsibility to
ensure that your final submission is free from Python or SQL syntax errors and that you follow
all instructions in this section.

We reserve the right to deduct points from your project if you do not follow the submission
instructions, or if you have syntax errors in your queries.

FAQ

Question:
I’m getting syntax errors when I submit to Gradescope, but I don’t see these syntax errors when I
run on BigQuery.

Answer:
Some things to check:
● Were queries copied correctly to the submission file?

8
 ● Are you using Standard SQL on BigQuery?
● Did you use backticks around table names?
Otherwise, this may be due to our autograder, which does not run with the same SQL that
BigQuery does. For example, our autograder does not support window functions (OVER,
PARTITION BY, etc.) and will throw an error. ​If you are sure that your query works
correctly (and is generalizable) with BigQuery SQL, then don’t worry​; we’ll be doing a
manual pass through submissions, and as long as your query is correct using Standard SQL
syntax then you will get full credit.

Question:
Do I have to match the column names given in the solutions? For example, in question 7, do I
have to name the column “upset_percentage”?

Answer:
You can name your columns whatever you want, as long as the content matches both in ordering
and in values.