SP18: Time Series Analysis

This is a review of the 2018 fall session. This course is a joke. The instructor is basically reading the notes, showing at the same time slides busy with formulas, which are never rigorously derived. The homeworks consist of yes/no questions that do not include any analysis problems that would require writing a code or trying to understand the learned material. The questions are of the type "If the mean of a time series depends on the time, then the sequence is stationary. Yes or No?" There is no feedback after the homework (at least in Audit mode): you only know whether the answer was yes or no.

The course breaks the edx enrollment clause "Audit this course for free and have complete access to all the course material, activities, tests, and forums." since some homework solutions seem to be available to "GT only" students. Moreover the solution to week 3 apparently has been released by mistake for all students, and the questions do not correspond at all to the questions in week 3 Audit mode.

There is no point to ask questions on the course forum since the instructors are absent (at least in Audit mode), and this fact is stated in the course syllabus. No way to discuss or obtain corrections to the course materials.
The final exam consists of multiple-choice questions where each question is repeated twice. It could be a mistake or an unusual way of grading.

The only thing good in this course is the availability of the R code and datasets discussed during the video lectures.

This course aims to be theoretical in a way since there's lots of somewhat complicated formulas with notations, yet little is done to derive these formulas for students. So it's like an unhappy medium; the course is neither rigorous theoretically or practical for those getting started with time series analysis. I thought about quitting the course earlier on in the semester, but decided to stick with it and I am glad I did. I did learn a few things that I had forgotten or not learned in my time series and forecasting courses during my degree. I audited the course, and completed all assignments and tests. I am very happy I didn't pay since I would have felt it was a waste of $99 (on EdX).

The bright spots of the course are the R examples. If you are looking to understand theory, it may be better to take another course or read a good textbook. If you are looking for practical time series analysis, try DataCamp and/or Rob Hyndman's book Forecasting: Principles and Practice.