Course: SML 201
Instructor: Michael Guerzhoy
S 2019

Description of Course Goals and Curriculum

Important points:

  • This course is a Data Science Course, not a statistics course or a programming course. You will learn both programming and theoretical statistics.
  • Introductory course appropriate for many disciplines.
  • Organized into learning R, data visualization using R, and statistical theory.
  • May feel fast-paced for someone with no programming background.
Overall, the aim of SML201 is to teach you to how to manipulate and analyze data using the language R. There is a lot of time spent on “data wrangling”, but that serves as the foundation of data analysis. SML201 pays a lot of attention to appropriate data visualization (graphs, charts, reports). Alongside learning how to code and how to use statistical tests, you will learn how to appropriately display your results so that the visualizations are not confusing or misleading. Since SML201 is an introductory course, it is appropriate for many disciplines. There is no background knowledge necessary. Professor Guerzhoy’s projects and assignments take on many different themes, from medicine to law, so that the course content feels relevant to any student’s prospective major. This course is structured somewhat differently from the Fall semester’s SML201, taught by Daisy Yan Huang. The first third to half of the course is oriented towards teaching how to use R. Later on, the course begins to transition into statistical concepts, where the emphasis begins to change from data manipulation to how to ethically and appropriately use statistical testing to evaluate data. This organization really hints at the heart of data science; as Professor Guerzhoy likes to say, data scientists are better at programming than statisticians and better at statistics than computer scientists. Though this course is intended for those without programming experience, it can feel fast paced for someone who’s never coded before. It is important to keep up in lecture and attend the (plentiful) office hours as soon as you are feeling behind. The readings should be used as a reference to help guide understanding of how to use R. As presented, the approach to coding in R is very different than popular languages like Python or Java, so even those with experience in these languages may be surprised by how R is used.

Learning From Classroom Instruction

Important points:

  • Course is divided into lecture to introduce material and precept to provide hands-on coding experience and guidance
  • Readings are optional but can be useful.
  • Precepts are hands-on coding sessions with assignments completed in pairs
The goals of this course are for students to come away with an understanding of the basics of R, data visualization, and statistics. More importantly, SML201 teaches students how to continue learning about these topics and how to use them beyond the context of the course. The purpose of the lectures is to introduce new material, and if Professor Guerzhoy feels the need to, review previous material that people are struggling on. Lectures are mandatory and involve lecture quizzes to ensure that students are paying attention. They can feel somewhat fast-paced, especially since the professor is a very skilled typist, but he usually posts the lecture code beforehand as a template that students can view. It is useful to download these files and use them as a template on which to make edits and comments to take notes on during the lecture. Additionally, do not be afraid to ask questions, request the professor to reexplain a topic, or to slow down a bit. He is very open to adjusting his lectures, and the questions help him know what topics need to be reviewed or re-taught. Outside of class, it is useful to review the lecture material to identify and address any confusion as soon as possible, to prevent falling behind. The readings are completely optional and are most useful as references for the topics covered in the lecture and precept. Professor Guerzhoy makes it clear that he will not test on material that he does not cover in lecture and emphasizes that the lecture and precept problems are what students should focus on in order to do well in the course. Precepts are essentially guided problem-solving sessions. There is a precept assignment posted each week, and they are designed to be completed entirely in precept. It is mandatory to work alongside a partner in precept, and preceptors will go around answering questions for anyone who gets stuck on an assignment. The precepts are excellent opportunities to clarify any confusion about lecture material. It would be useful to review the previous week’s lecture material in advance of a precept, but not entirely necessary. If assignments are not completed in precept, they can be completed at home.

Learning For and From Assignments

Important points:

  • Weekly precept problem sets reinforce lecture material
  • There are three projects and two midterms
  • All assignments and assessments test applications of lecture material.
As mentioned before, the precept problem sets are done in precept and are designed to reinforce the topics learned in lecture. Though the lecture material may seem straight forward at times, actually applying the material in precept can be more challenging than expected. Successful completion of the problem sets ensures that you are keeping up with the course. Understanding the methods of solving the problem sets well is also important to success in projects and the assignments. There are three projects assigned throughout the course, each one focusing primarily on the lecture material immediately preceding it, but also building upon one another. The projects involve data manipulation, data visualization, and critical thinking in order to analyze the data and draw appropriate conclusions, and are entirely written in R. They can be viewed as “advanced problem sets” since they are organized in a similar fashion. Projects can be completed as pairs or individually. The professor makes available an abundance of office hours, especially as the due date approaches for these projects. It is best to work on these projects gradually rather than all at once, since it is easy to run into bugs or errors while coding that may require a preceptor’s help to address. There is ample time to complete the projects. There are two “midterms” in the course, done-in-class. The best way to prepare for these midterms is to ensure a strong grasp of the materials covered in the problem sets or projects, since the questions address the same concepts. It is also necessary to understand the conceptual knowledge behind the topics covered in lecture; do not just memorize the code. Furthermore, since the exams are on paper, you should become comfortable with coding on paper in advance of an exam. These exams are generally not a huge portion of the course grade, so it is possible to recover from an exam that did not go well.

External Resources

Important points:

  • Many office hours available
  • Make appointments with preceptors/professor if needed
  • Piazza and Google are your friend
There are plenty of resources available for this course. McGraw has some general R tutorials available, but preceptor office hours should be considered the primary source of help. Come prepared with questions regarding things to clarify, or any bugs encountered in the code. Professor Guerzhoy is very flexible and often available for any questions, you can stop by his office or make an appointment with him. HE also holds review sessions occasionally on the weekends, that are basically mini-lectures where he re-teaches material from his lectures based on the topics that people request. There is also a course Piazza page that is set up, where professor Guerzhoy will regularly post tips and announcements. He also has an extremely impressive response time on Piazza (avg 4 minutes!) to any questions that students post, even answering questions in the evenings. Don’t be afraid to reach out to him if you need help. Finally, Google can be a useful resource as well, for learning specific things that you may want to do with your R code (such as rearranging a graph). This is completely optional, since all the necessary material is covered in lecture already. Any code that you use after learning from online should be clearly cited.

What Students Should Know About This Course For Purposes Of Course Selection

Important points:

  • Learn how to use R, how to draw appropriate conclusions from the data, and ethical use of data analysis
  • Skills applicable to many fields; data analysis occurs in research, medicine, social sciences, computer science, natural sciences
  • Workload ranges from level of experience with coding and statistics; more experience makes this course less time intensive. Roughly 2-10 hours a week outside of class (depending on precept problem set and project, as well as experience)
  • Most beneficial things: learn more about the nuance involved in statistical analysis, and both the strengths and limitations of this tool
  • This course can be pretty manageable but requires a decent amount of dedication to the subject. The professor is very approachable and flexible.
SML201 is a very useful course that teaches you how to appropriately analyze and draw conclusions from data. It allows you to pursue work that involves analyzing large sets of data; a pretty common occurrence in any research field. Furthermore, the course teaches you how to critically think about data and how to draw appropriate conclusions from statistical analyses. This makes the course relatively popular with students of all disciplines, including pre-meds, engineers and social science majors. Prior programming knowledge is relatively helpful for this course, since it makes the understanding of some of the syntactical structure of R easier. This is useful but not necessary to do well in the course. Prior statistical knowledge can be useful as well. Depending on one’s level of background, and whether or not a project has been assigned and the level of difficulty of the precept problem set, the workload for the course is from 2-8 hours a week outside of class. Keep up with the assignments and you should be fine. One of the most valuable things learned in this course is the limitations of statistical analysis, and its ethical applications. Statistical analyses are often misunderstood and can be poorly applied. At best, this leads to accidentally misleading conclusions from research, while at worst statistical analyses can be used to justify social inequalities. The course demonstrates that data analysis can only be as good as the data being used; and will reflect any biases already present in the data. Prospective students should know that SML201 is a pretty manageable course that requires a decent amount of dedication to the subject. The professor is very approachable and flexible, and does his best to ensure that everyone in the course will succeed. He will work with you to make sure that you will do well. Any student will benefit from the knowledge gained in SML201.
Introduction to Data Science

Add a Strategy or Tip