CSE-6040 - Computing For Data Analytics

Practice problems will release along with Notebook 12 on October 21st.

The practice problems released are all past exams, except for the extra example problem. They are all accessible through your LMS and hosted in Vocareum like your notebooks. The assignment titled "Practice Problems for Midterm 2 - Untimed" contains all of the past exams we're sharing with you. There is also a "Practice Timed Exam for Midterm 2" which is the most recent past exam, and is setup like a test in Vocareum (with time limit), without the proctoring. Note: as you access that assignment through your LMS, there's a note about how we recommend using it and a couple of "watch outs".

The information listed below gives you a quick intro to the notebooks, and where available it gives you the exact point cap and time limit from when we offered it.

Note per pinned Piazza post (@815 for GT, @146 for VMM): The tests below are representative of the problems, but there is minor change you will see on the exam: some problems will specify that you use either Pandas, or SQL. Previous exams allowed both (with varying levels of difficulty), but this new requirement means you will need to focus on both in order to answer all of the exercises. See PMT2-EX1 below for a quick example.

Exam prep tips:

Suggestion 1: while we're releasing all of the solutions with the notebooks, we recommend you don't use those until you finish the whole practice exam, or if you get stumped on an exercise for an extended period of time. Learning how to resolve and hunt down those issues on your own is the skill you need to build in order to prepare for the exam.

Suggestion 2: You're free to discuss these practice exams openly and share your code with classmates. Connecting with others on Piazza may help you think through alternative approaches to problems.

Special Example Exam: PMT2-EX1(MT2 SQLite and Pandas) Solution

• Note: this is a quick demo created to show you what Pandas or SQL-only exercises look like
• Topic: Star Wars
• Key skills: SQL, Pandas

PMT2-SP24 Solution

• 7 exercises; 13 available points; 7 points required for 100%; Time limit 4 hours
• Topic: NYC Roadmap - Traffic fatality density. In this notebook you will look through some NYC traffic data to build a KDE model
• Key skills: SQL, Pandas, Numpy

PMT2-FA23 Solution

• 9 exercises; 21 available points; 13 points required for 100%; Time limit 4 hours
• Topic: "Punt, Kick, or Go for it?". In this notebook you will look through some NFL play results and create a risk/reward model.
• Key skills: Pandas, Numpy

PMT2-SP23 Solution

!!Accessible through standalone assignment listed after the full "Practice Problems for Midterm 2 - Untimed" assignment

• 11 exercises; 21 available points; 12 points required for 100%; Time limit 4 hours
• Topic: "Better Reads". In this notebook you will look through user-generated book reviews to uncover "communities of users".
• Key skills: SQL, Pandas, Numpy

PMT2-FA22 Solution

• 9 exercises; 19 available points; 12 points required for 100%; Time limit 4 hours
• Topic: Capturing Data Changes for Slowly Changing Dimensions. In this notebook you will implement a common data engineering paridigm to maintain a historical record of some mock-up business data.
• Key skills: Manipulating tabular data with Pandas and SQL, Strings
• Note: Most exercises in this notebook are solvable with Pandas or SQLite. However, the solutions using SQLite need to be "massaged" to pass all of the test cells. Pandas is strongly recommended.

PMT2-SP22 Solution

• 9 exercises; 17 available points; 12 points required for 100% (lowered from 14); Time limit 4 hours
• Topic: Actor network analysis. In this notebook you will explore a dataset of film credits and create/analyze a relationship network of actors starring in the films.
• Key skills: Pandas, native Python data structures, and incorporating new tools given appropriate documentation.

PMT2-FA21 Solution

• 7 exercises; 13 available points; 9 points required for 100%; Time limit 4 hours
• Topic: Campaign finance geography. In this notebook you will calculate how similar ZIP codes in the United States are to one another based their residents' donations to political candidates in the 2020 election cycle.
• Key skills: Pandas, SQLite, sparse matrices