Back

Data Science for Engineers

Credits: 4

Tags: Data science, Python

Class Overview

This course is an introduction to data science for engineers. The data science workflow: acquisition and cleansing, exploration and modeling, prediction and decision making, visualization and presentation. Tools for data science including numerical optimization, the Discrete Fourier Transform, Principal Component Analysis, and probability with a focus on statistical inference and correlation methods. Techniques for different steps in the workflow including outlier detection, filtering, regression, classification, and techniques for avoiding overfitting. Methods for combining domain-agnostic data analysis tools with the types of domain-specific knowledge that are common in engineering. Ethical considerations. Optional topics include classification via neural networks, outlier detection, and Markov chains. Programming projects are in Python.

Prerequisites: MATH 1920 and either CS 1110 or CS 1112. Corequisite: MATH 2940
Key Topics: Numerican Optimization, Fourier Transforms, Principal Component Analysis, Image Processing, Probability, Markov Chains

Professor: Dr. Jayadev Acharya

Semester(s): Fall, Spring

Difficulty: 4/5

Rating: 3/5

Assignments: Weekly problem sets, and biweekly labs and pre-lab assignments.

Exams: Two prelims and one final exam.

Pros

  • Practical projects that apply theoretical knowledge.
  • Well-structured lectures and clear explanations of complex concepts.
  • Comprehensive and organized course materials.

Cons

  • Heavy workload that requires significant time investment each week.
  • Challenging exams that require deep understanding.

Tips for Success

  • Make sure to review lectures regularly.
  • Start working on the projects early; they take more time than expected.
  • Attend office hours and study groups to tackle difficult topics.

Professor: Dr. Vikram Krishnamurthy

Semester(s): Spring

Difficulty: 2.5/5

Rating: 4.5/5

Assignments: Problem sets (involve math and Python programming for simulating)

Exams: 10 minute quiz at the beginning of almost every lecture, one take-home final

Pros

  • Quizzes are very fair and help you stay on top of the material
  • Well-structured lectures and some overlap (especially for probability and random variable section) with classes like CS2800 and ECE 3100
  • Professor cares about students and designs class to not be too stressful

Cons

  • Need to stay on top of it for quizzes
  • Material can be repetitive depending on what other classes you have taken

Tips for Success

  • Office hours are usually not very packed (especially towards the start of an assignment's release date), so take advantage of them
  • Knowledge of linear algebra (eigenvectors and eigenvalues) are helpful for understanding the material, but professor will re-teach it from first principles during lecture