Course Details

Pros: 1. Easy A or A- at worst. 2. Does cover many fundamentals from Ordinary Least Squares (classic multiple regression) to Bayesian Statistics to Other Topics including PCA Theory, Fixed Effects, etc. 3. Just 12 assignments (2 part each) and no exams. Can drop 2 lowest assignments. They are all peer-reviewed, though. Cons: 1. Boring and recorded lectures don't feel engaging. One of the worst in the program in terms of this matter. 2. Organization of topics is not necessarily coherent at times. Could benefit from having suggested textbooks even if the instructor claims, topics are "self-contained" 3. No effort by the course to use realistic datasets. R scripts are disorganized and use weird notations, variable names like xxx instead of x^3. Detailed Review: Covers fundamental topics in regression. Many people dislike this class, for reasons I mentioned above under cons. But at least it is straightforward to take, then move on afterwards. For a supplementary reference, I recommend Regression and Other Stories by Gelman et al., although this book is really for DSC 384 (Causal Inference). The first half is also a suitable text for DSC 382.

Pros: 1. Reasonable workload, including for summer. Predictable format week to week. 2. No exams, drop lowest 2 homework grades 3. Professor and TA fairly responsive Cons: 1. Lectures often refer to data without any context 2. Starter code uses minimalistic naming with poor formatting and most files have no comments 3. No textbook Detailed Review: I suppose if there were any silver lining to the class, I was able to sharpen my math skills (including typesetting) a bit given that I don’t have the strongest mathematical background. If you’re in the same boat, I would recommend taking or reviewing probability beforehand, as opposed to taking probability and regression concurrently. This course specifically applies probability distributions, expectation/variance, and maximum likelihood estimators throughout the semester, and at the very least, I was able to solidify some concepts that I wasn’t completely solid on when I took DSC 381. Prof. Walker was fairly active on the Ed discussion board. The TA and the LF’s were pretty helpful and quick to address issues, including with the peer grading software Peerceptive (which the grad program appears to have scrapped, thankfully). However, the discussion board was a ghost town by the end of the semester, other than when there were grading issues. Overall, I didn’t really find the class enriching, especially in the broader context of this program, and my thoughts aren’t too much different from most reviews here. The material isn’t presented in a very interesting or applicable way. The lectures and starter code especially are bare bones. I suppose this class is technically “self-contained”, though the content is presented in such an half-baked, sometimes inscrutable manner that people frequently asked for outside resources. Thankfully, a lot of the topics are covered in other courses. I’m taking machine learning right now, and the two professors teach linear/logistic regression, MLE, Bayesian inference, and EM with so much more clarity.

Pros: 1. Not that hard Cons: 1. Miserable slides 2. Miserable code 3. Miserable Detailed Review: Stunning to me that UT still has this course up in its current state. Going into this course I was excited for a theoretical deep dive with some potential applications for regression. That's not what it is. This course is heavily theoretical and poorly explained. I spent more time reading outside resources than trying to struggle through Dr. Walker's explanations and slides. His attitude online was also not that great, it almost seems like he expects everyone to already have a PhD in stats? I think the most frustrating aspect was the numerous instances where the professor simply hand waived an important part of a derivation or said something to the effect of, "well everyone knows this so I'm not going to take the time to explain this in the slightest." In the same time he could have simply explained it. Working through "example" in class was also frustrating. There was no working through it. Just vomiting of stats associated with a data set which is not provided meaning you can't actually replicate the work yourself. Also, the code is garbage. The only redeeming part of it is the grading system makes it relatively easy to do well. You won't learn much, but you will do well.

Pros: 1. TA’s were helpful where they could be 2. Lectures are (generally) independent of one another and no final exam Cons: 1. This class felt like it was taught to the wrong audience 2. Non-interactive lessons 3. No textbook / lack of great materials Detailed Review Pros: 1. The TA’s feel like the glue holding this course together. They helped to navigate the bad grade curving of Peerceptiv and were very responsive to questions. 10/10 for them 2. There are 11 modules, all of which aren’t really dependent on one another, and there is no final exam. Thus, you don’t have to remember much from unit to unit in order to perform well. This made it nice while trying to navigate this black hole of a course over time. Cons: 1. Many times in his lectures, Professor Walker would say things to the effect of “I know you all have a good idea of [X topic] so I won’t go into it too much here” or “[Method Y] is something I’m sure you know” when, in fact, most of us don’t know what [X topic] or [Method Y] are at all. It felt like Professor Walker was teaching to the wrong audience, as if he was expecting all of us to have a specific set of precursor knowledge. That led to a lot of confusion and disconnect with the lectures. 2. This whole course was Professor Walker talking through PowerPoint slides. Nothing more. He didn’t even have a mouse cursor so we could follow what point of the slide he was talking about. This made lectures hard to follow and boring. Sometimes he would have some example “experiments” he coded out so that we could see how various statistical tests worked, but there was no expectation of us to “code along” or to think critically. Just hours on hours of listening to him speak. I ended up skipping most of the lectures toward the end of the course. 3. There is no textbook for this course and the lecture slides are very lecture-specific, meaning you can’t just read the lecture slides by yourself and get the whole picture. Many of the slides have edits to them that are only mentioned in the lecture (assuming the data is centered around the x axis, assuming you only have one feature, etc) such that the slides aren’t useful outside of this class. On top of that, there is R code provided alongside the lectures, but the R code is some of the worst code I’ve ever seen. It isn’t documented well and is littered with inefficient code and bad variable naming. It's hard to learn when there are no good materials to pull from... Ultimately, I felt this whole course was interesting but won’t generalize well outside of this limited environment of a class. I can definitely see why this is the lowest rated class in the program and hope this class can be revamped in the near future!

Pros: 1. Good curriculum: a lot of important concepts are covered. 2. 3. Cons: 1. Terrible grading system 2. 3. Detailed Review: I didn't have as much of an issue with the course curriculum as other reviewers have had. I thought Professor Walker was knowledgeable, quick to respond to messages, and overall a good professor even if his lectures were a little dry and his lectures note in need of a revamping. I also didn't mind that there wasn't an application part of this course. In my opinion, this course serves as the theoretical foundation that you'll be using for other courses which do have practical applications. My least favorite part of this course is the grading system. There is going to be variance in a student's grade (sometimes you just have a bad day), but this course felt to me to be one of the only courses where a student can be performing at the 75th percentile in terms of understanding the material but be scoring at the 25th percentile. It's fine when most of the class gets As, but there was (justifiably) pandemonium when an assignment was originally set to be curved to a B. Let me expand on this further. There are 6 long form written answers homework assignments which are graded by fellow students. I think the grading is inconsistent at best, and nefarious in the worst case since grades are curved. Sometimes students will grade you with their own interpretation of the question which is not correct. Sometimes students won't grade you correctly if you had an alternative answer which is also correct if it doesn't exactly match the answer key. It is my belief that the highest scoring students are those who attend office hours and get the answers directly from the TA (since the answers provided will match the answer provided in the solution key). It is also my belief that academic integrity is an huge issue in this course (although I don't care to elaborate on this further). All this to say, if this course was ever curved, it would not be a fun experience.

Pros: 1. Learned new things 2. Could drop 2 assignments 3. Good to start with R coding. Cons: 1. Recorded classes were not great 2. Needed to do a lot of research to understand some assignments 3. No textbook Detailed Review: As there is no textbook, finding material with the same approach as the professor was not easy. I had to look for a lot of information before completing some assignments. If you have a recent background in Statistics and Math, probably you can do well in this class.

Pros: 1. Easy A 2. You know about most important regression techniques after this 3. Cons: 1. Terrible teaching 2. Terrible slides 3. Why am I paying for this? Detailed Review: This class is garbage. Let's start with the material. The material goes over most regression techniques you should care about. This is good. However, how everything is taught is awful. Prof basically reads off the slides. I stopped watching lectures in week 2 and just paid attention to the course materials. Therefore, it's just a self-study, right? That would be fine if the course materials weren't also complete garbage. Let's start with the quality of the slides. They are extremely disorganized, sometimes making logical jumps that require a lot of thought to understand. Then, on the other hand, the slides sometimes oversimplify everything to the point where you don't understand what is actually happening. An easy example is the week 1 slides that cover simple linear regression (without matrices). After introducing the topic there is a whole aside to sample estimates that isn't explained very well, THEN they talk about least squares (with a loss function that isn't even the MSE! I know the constant goes away in the derivative but why not include it???). The least squares explanation is done pretty well and then he mentions that you can do the same thing with MLE but doesn't really explain the specifics (what is the pdf???), just some basics about MLE that you should already know. If he's going to explain things we should know, then do that for everything instead of making crazy logical jumps at some points. Then, he says he'll go through the math but does it for a simplified case (\bar{x} = 0) that nobody would care about in real life. The not-simplified case is literally not much worse, just an extra covariance term. Why not do that? He also generally simplifies for \sigma not known which is understandable but for god's sake just give the results for \sigma not known in a slide or something instead so someone who's interested and wants to apply this in real life doesn't have to spend 20 minutes deriving it or going and googling. Then, there are so many mistakes/misleading things I've caught in the slides that are ridiculous. An example is the discussion of \hat{\sigma}^2/\sigma^2 following a chi squared. The probabilty given is not enough to figure that out since it would require a ton of assumptions to even be remotely valid. The result given is implied to be the distribution of \hat{\sigma}^2/\sigma^2 but if you actually read carefully it's not: it's the result for the RSS^2/sigma^2. Misleading. He also states the wrong thing for the variance of the residuals. Then, he doesn't mention the distribution, making you assume it's normal like most other things in that lecture but it's ACTUALLY T. It's not just me catching it: other people were posting about mistakes they found on Ed well into week 12 of the course. These examples I gave were just the most glaring from week 1. Absolutely ridiculous. Wouldn't be surprised that there is no textbook for this course for the sole purpose that more people would catch slide mistakes by reading the book. I get that this is a theory course, but they can at least mention how you should actually do this in practice. The code given is all done with matrix operations and iterative procedures. Sure, make us do it like that on the HW to learn the theory but then give us another file showing us how to do it in practice once the HW is done. I should not be having to do crazy things to run a cox proportional hazards model in real life when I can just use a survival library. And let me tell you about this man's code. It is genuinely the worst code I've ever seen. I would have more faith in an AP Comp Sci A student. It's not even about efficiency although that can be improved: it's the formatting. There are basically no comments and everything has the most garbage names that you can imagine. Variables are usually one or two letters that have no relation to what they represent. I got so fed up with this that I started rewriting all the code by myself by week 4 which was far more time-efficient than parsing his. At least it's an easy A. There are 12 HWs and the top 10 are the entire grade. No exams or quizzes or anything. Most homeworks are quite easy (exceptions I say would be HWs 5, 10, 11) and doable in an hour or two if you have any sorts of a math/stat background. Most HWs are peer graded which allows for leniency with silly mistakes as well. The TA when I took this was great and active on Ed which helped clear up any confusion that may stem from mistakes in the slides. Prof was active for the first week or two then disappeared other than questions about grades (obviously it won't be curved etc) It's an absolute scam that I am paying for this course; I feel like I was the victim of a mugging. I could literally make a better course from scratch in a week and teach it better. I thought the probability course could've been significantly better but it was at least worth a hundred bucks. I should be paid for having to take this course. If you actually want to learn regression without countless hours of deciphering slides and individual research, go to MIT OCW or YouTube: it's FREE!!!

Pros: 1. Low-effort credit for the summer 2. N/A 3. N/A Cons: 1. The lectures are not helpful in the slightest 2. The slides/supplemental materials and sample code are indecipherable 3. There is zero practical application during the course Detailed Review: Others have summarized the course far more effectively than I could hope to. I will just say this: the course is as bad as the reviews indicate. Take it in the summer, grit your teeth, get an A, and move on. Search on YouTube for instructors who put more effort into free online content than this paid course.

Pros: 1. Not overly difficult (except HW5) 2. Reasonable summer time commitment 3. Cons: 1. "It's not overly complicated..." - lectures are endless recitations of equations with no context 2. Professor actively refused to recommend any outside resources to learn the material 3. www[tt] <- fff - the sample code will make your eyes bleed 4. Non-standard terminology makes it difficult to cross-reference outside materials (Bernoulli regression anyone?) 5. I got an A, but I didn't learn anything 6. Only one TA for well over 100 students (no complaints on him personally, just need more) Detailed Review: I went into this class thinking there was no way it could be as terrible as the other reviews say. Unfortunately, dear reader, I have bad news for you - it is indeed as terrible as the other reviews say. I'm not going to repeat the same topics as all of the other reviewers, except to say that this *should* be an interesting class. This *should* be a relevant class. There is a lot here that is foundational to data science and machine learning. But the professor does such a poor job of communicating it that most opportunity for learning is lost. It's not just the worst-taught class in the program, it's the worst-taught class I've ever taken. Protip - don't bother with HW5. Just use one of your two drops and move on.

Pros: 1. No exams 2. Not bad to take during summer 3. Cons: 1. No real world application whatsoever 2. The worst sample code you could possibly imagine 3. Detailed Review: I can't remember a class that I took that I took less away from. I can confidently say even after getting an A in this course, I don't know any more about regression now than I knew going into this course. There are no practical problems assigned. There are no practical applications discussed in lectures. I gave up even watching them after the first week, because they are just reading directly off of the slides most of the time. It is very disappointing that this is considered a "foundational" course. I took absolutely nothing away from it. The sample code provided is a total joke. Just really disappointing all around.

Pros: 1. 5 multiple choice homeworks to avoid all peer grading. 2. 2 homeworks dropped 3. There are no exams. Cons: 1. The code. Holy Swiss cheese, Batman! I would have failed any CS class where this was my output. Completely indecipherable. 2. No application to real world data sets. 3. Not sure I feel confident in applying knowledge from the course to a data science problem and isn't that the point? Detailed Review: What I hope to gain from this masters is an understanding of the application of mathematics to data science issues. This course is not effective at addressing real world data science wrt regression. I understand the algorithms that can be applied but I don't know when, why, and in many cases how as we never explored the types of data that these algorithms would apply to. The data generated was often purely random based on a generic distribution. As a data scientist- I feel like I should also be aware of issues that arise while evaluating a data set. There are so many data sets available out there that apply to these topics that could have been used to discuss these topics or to compare approaches but the code given was so obscure and difficult to read that I generally ignored it as it would take hours to decipher. You can throw the provided code into ChatGPT and ask it to give you notes on each line which helped to understand what had been coded. (Even the AI was throwing shade about how bad the code was- pointing out inefficient uses of nested loops, extraneous variables that were unused, etc.). The data sets used in problems were meaningless. Why? This topic could be so fascinating if we were gleaning insights from actual data sets- even working the exact same problems. The lectures also simplified derivations to the point that you could do large sections by hand by making assumptions about the data- convenient for the course questions, but not additive career wise as data encountered outside of a pure mathematics distribution would like have pitfalls or issues that I was hoping the instructor could lend their insight to: how to address, how to spot issues, etc. I don't feel well prepared after this course to apply regression concepts to real world data. I'll have to take another outside course to understand this topic from a data science perspective.

Pros: 1. Very easy course to score highly in 2. The HW problems that involved data analysis were interesting. 3. Cons: 1. The lectures are terrible. (Can I list this for Cons 1, 2 and 3?) 2. 3. Detailed Review: The course is definitely the worst of the 4 classes that I have taken so far in the program. The lectures are highly theoretical and it seems like the instructor is just rambling and reading directly off of the slides. The course grade is based only on homework, so it is easy to get a high score in the class with little effort. The HWs definitely varied in difficulty and utility. Some were highly theoretical (I didn't get much out of these), while others had data analysis portions that were useful. I would definitely recommend taking Probability before this course as it builds on some of the concepts from that class.

Pros: 1. The course material is well organized. 2. There are only homework assignments, no tests. 3. It is definitely possible to get an A with sufficient effort. Cons: 1. I learned almost nothing of value. 2. The lecture videos are unbelievably dry and boring. 3. I don't feel confident applying the material to a context outside of this class. Detailed Review: This was one of the worst courses I have ever taken as a student. Although the course was well organized and it was possible to receive an A in the course with sufficient effort, I learned almost nothing of value. The lecture material focuses far too much on derivations and not enough on the background information that provides context for the material and motivation for why the students should learn the material in the first place. I do not feel confident applying the information I learned in a context outside of the narrow confines of this course. Dr. Walker is clearly very knowledgeable on the subject matter but he does not do a good job of communicating information in a way that is engaging or motivating for students. His teaching style is very dry and he mainly just repeats what is on the lecture slides without providing much useful, additional information. In fact, I stopped watching the lecture videos after the first couple of weeks because I realized that my time was better spent just directly reading the lecture slides. I wish he used the lecture videos as an opportunity to provide context for the material on the lecture slides and to provide examples of how this information could be applied in a practical setting. The TAs for this course were incredibly helpful. Their responses on Piazza clarified a lot of misunderstandings and were essential for understanding the lecture material and completing the homework assignments.

Pros: 1. 12 homeworks with 2 drops, no exams 2. Covers a wide array of critical topics Cons: 1. Lecture materials and their delivery are subpar 2. Most lectures examples are not reproducible, given "code" is borderline unreadable, and there is no textbook or recommended reference materials 3. Homeworks are heavily theoretical and do little to expand beyond lecture 4. Professor's attitude in lecture and on Piazza was not conducive to learning Detailed Review: Despite the consistent negative reviews this class has received in the past, I hoped that it would be a worthwhile experience after reading the list of topics covered in the syllabus. I was sorely mistaken. This class was week after week of the professor barreling through the material and heavy amount of math while calling the majority of the material "trivial" or "easy", often referring us to a nonexistent webpage for more details. Despite this course being offered since as early as 2021, the number of typos remaining in the lectures was unacceptable. Luckily, the TA was very on top of fixing these issues and responsive to the questions and concerns students had over their understanding of the material. The workload of this class comprises of 12 homeworks, roughly half peer-graded half multiple-choice machine graded, with 2 drops. The homeworks are always two 5 part questions, which sometimes conflicted with the lectures which were divided into 3 segments. In weeks where there were 3 major topics, one of them would just not be on the homework at all. The overwhelming majority of the homeworks consist of deriving various theoretical properties about the material that was covered, expanding little beyond them since it is often just a simpler version of what was covered in the slides (single predictor vs. multiple predictors for example). Derivations that probably should have been included in the lecture material simply end up in the homework. Sometimes there is data given to "apply" the concepts too, but the data is just "x" and "y" with no real-world connection so those problems devolve into mindlessly putting the data through the equations. The professor uses R in the lecture slides, but having a rigorous grasp of it is unnecessary. The provided code examples are near unusable due to poor variable naming and no comments. You are better off writing the code from scratch, most of it is just applying a formula or implementing simple algorithms anyway. "Examples" in the lectures are also almost always based on randomly generated data, so it is impossible to reproduce the exact values from the slides making the few examples included to begin with near useless. Overall, this class covers a lot of important and interesting material, but don't worry too much about learning it deeply here. Nearly all of the topics were covered in other classes in the program in a more useful way, so this class mostly just became a chore to do each week. With some effort, succeeding in this class is pretty simple, but it is unlikely that this class alone will help in a data science career. Hopefully, changes will be made to this class one day, but given its history I highly doubt that will happen. PS: Forgot to mention this but if you haven't taken DSC 381 or another introductory class to probability recently, make sure to brush up on that material because a grasp on the basics of probability distributions and their properties is expected but never explicitly reviewed. In particular, the properties of expectation and variance of random variables is something that comes up on nearly every homework (and something that people consistently struggled with every homework).

Pros: 1. Bayesian distributions 2. Gibbs sampling Cons: 1. Nothing pactical 2. Theory is in areas you're unlikely to encounter Detailed Review: If there was a class ripe for overhaul, this is it. I want to compare it to NLP at a meta level, because NLP is not really the same thing but was expertly delivered. In NLP, we touched on the theory behind language, pytorch (I hadn't done DL yet), and several approaches to generating text including LSTM and transformers. We still managed to do practical projects that I'm proud to have completed. In this class we seem to have endlessly computed conditional distributions, and did a little bit of linear algebra that might be important to be aware of but I don't think so. There are fundamental aspects of practical regression modelling that were not touched on at all. There was no mention at all of the assumptions of ordinary linear regression or how to test for them. I would bet that the majority of the class's don't know what heteroskedasticity or serial correlation even are, let alone what an approproate test for them might be. We went really deep into a bunch of Bayesian stuff. I thought the math was quite interesting from an intellectual point of view, but I have no idea why you would ever do it in practise. I've done more than a few regressions in my life, but I've never approached them with an idea of a prior mean and standard deviation for the parameters. I don't know when I would ever find myself in the position of understanding the parameters so deeply but sill needing to do a regression. If I was at that point, surely I would move onto ML methods. I've seen complaints on here about classes that were just very hard, but still quite good. This is not that, it's one that I wish I could have avoided. Not worth the money or time I spent on it.

Pros: 1. lots of math, and help build the intuition and linkage with various theorem 2. good opportunity to build stats functions like t-stats from scratch 3. homework only, no exams!! grading is also reasonable can drop 2 homework Cons: 1. no practical examples 2. R code is hard to read with useless coding style xx yy zz 3. the EM Algorithm could go much further but only covered in the last 2 weeks Detailed Review:

Pros: 1. Homeworks are very simple once you look up relevant resources and understand what it is trying to make you do 2. Doesn't require a lot of time commitment 3. Opportunity to flex your math muscles (e.g. doing derivatives) 4. Contents are pretty broad 5. Instructor and TAs are very nice Cons: 1. I passed with an A without gaining a good understanding of anything Detailed Review: I feel it is a difficult course. I had difficulty understanding 90% of the materials. Then when I started on the homework (especially the later ones), I didn't know where to begin. However, when I read up on the online resources, I realized that many of the problems actually do have well-structured solutions (i.e. step-by-step methods that you can follow and apply to solve similar problems). Unfortunately, those structured solutions were not presented in the course materials. The course materials lacked examples. In the scarce examples that were presented, data was randomly generated without seed and critical steps were omitted from the process (sometimes only intermediate results were provided), making it difficult to replicate and check your understanding. In some cases, more practical approaches (such as one-line R code) exist, but were never mentioned in the course materials. R/Python was only used as a "calculator", which is probably a reason why the provided R code is difficult to interpret. I understand if the purpose is trying to make us dig under the hood. However, I feel from a practical point of view, at least knowing what pre-existing functions we can use will be very helpful.

Pros: 1. Lots of content is covered 2. No exams 3. Fairly straightforward homework (not too many hours to complete) Cons: 1. Example code needs to be refactored 2. No reference textbooks on syllabus 3. Some of the math notation on the slides is confusing Detailed Review: Overall I really enjoyed all the contents of this course. It did cover material that you would normally find in a regression course like simple / multiple linear regression, regularization, ANOVA testing, F-tests, and logistic regression. However it also went into several other more advanced topics such as Bayesian regression, nonlinear regression / newton raphson, parametric and nonparametric bootstrapping, GLMs, metropolis hasting algorithm, Gibbs sampler, EM algorithm, mixture models, and non-parametric methods such as trees, splines, and kernel smoothing. The only downside is that most of the methods weren't put into context of an example problem / dataset. So you will need to find external texts / references such as Elements of Statistical Learning for that. The homework for the most parts were fairly straight forward. Most took anywhere from 2 hours for the easiest assignments up to 8 hours for the toughest ones. Around 40% of the assignments are multiple choice and the rest are peer graded. Most assignments were around 60% deriving various mathematical results and 40% programming (most used R, but you are free to use whatever programming language you prefer). All of the datasets are fabricated/simulated data made up from the course staff, so again no "real world" toy datasets to apply the material with. You pretty much have to code up the solution using only basic tools such as matrix operations, functions, and loops to derive the solutions and not use pre-made packages. I thought this was good because it reinforces the mathematical theory even though you would never use these implements in a real world setting. The example code needs to be refactored. Plenty others have complained already about this so I won't go any further on that. Some of the notation in lectures is ambiguous. For example vector and scalars don't use the typically notation where the vectors are differentiated from scalars using bold font. Also the prime symbol ' can mean a matrix transpose or a derivative. Sometimes a random variable would be written as equal to its pdf function instead of showing that it's distributed according to that distribution. You could usually figure it out most the times by the contexts, but it did make following the slide more difficult than necessary at times. Overall I enjoyed the course.

Pros: 1. Homework nicely complement the lectures. Doing the homework greatly improved my understanding of the topic. 2. Some exposure to Bayesian statistics. 3. A deep dive into regression will help you become a more well-rounded data scientist. Cons: 1. No textbook to refer to -- you'll have to go just off the slides and lectures. The instructor says the course is "self-contained" but it really isn't. 2. The example code is just plain bad -- no comments, non descriptive variable names, etc. Detailed Review: I enjoyed this course and I now have a greater appreciation for how deep a topic regression really is. There was a good mix of theory (although there weren't any proofs per se, the mathematics was very much a central theme throughout the course) and application (you will implement Newtons method a few times, you'll be exposed to EM algorithm, you'll work with R, etc). This is not a course where you blindly import some Python/R package and run "fit" on a data set and have no idea what's going on under the hood. Although I was happy to get some exposure to Bayesian statistics I think there should be a Bayesian Statistics course pre-requirement. Not because the Bayesian sections were particularly difficult or anything but because it felt like Bayes was kind of just "tacked-on" when really it deserve much more attention than this course is able to provide.

Pros: 1. Light workload 2. Can be taken in the summer 3. Can say you took a course on regression models 4. Other students are really good / sharp. 5. The course is effectively only 9 weeks long, since you can get a numeric grade of 90 by completing the first 9 assignments with a grade of 100, and letting the assignments for weeks 10 - 12 go (90 is the cutoff for an 'A'). A numeric grade of 78 or above is the threshold for a letter grade of 'B', which is even easier to accomplish by completing all assignments through week 9. Cons: 1. Get an easy A, but learn nothing useful 2. Assignments generally due on Friday evenings (makes zero sense for working professionals) 3. Course staff not very helpful with questionable professional scruples 4. The course is effectively only 9 weeks long, since you can get a numeric grade of 90 by completing the first 9 assignments with a grade of 100, and letting the assignments for weeks 10 - 12 go (90 is the cutoff for an 'A'). A numeric grade of 78 or above is the threshold for a letter grade of 'B', which is even easier to accomplish by completing all assignments through week 9. I think the biggest issue for me with this course is that I'm not exactly sure what it's intended to accomplish. It doesn't dive deeply enough into the theory to provide a really solid foundation for understanding this material on an abstract, theoretical level (case in point - the week of material on Bayesian statistics might be the worst presentation of this material I've so far come across - shallow, superficial coverage which isn't going to see you learn anything useful), and the applications are either lacking depth and context or - in some cases - just nonexistent. So I'm not sure that one should even feel confident applying these methods in real-world scenarios procedurally / mechanically to solve actual problems. Overall, it feels like someone said "Oh, this course of study should have a regression class", and then hastily threw one together without giving much thought to what would be of actual value to folks hoping to work as data scientists; it's not even entirely clear that the course creators really know what would be of value to DS people, frankly. The code samples the course provides are embarrassing, and you definitely wouldn't want a prospective employer to catch you coding R the way these people do. A for loop? In R? Really? I think that what I'm noticing is that there's a significant gap between what academics think DS is, and what industry thinks it is. And the issue there is, it's not the academics that are doing the hiring for these roles. It's industry. ********** PRO TIPS: ********** Don't watch the lectures (they're worthless). You'd just be wasting your time. Skim the slides, and then go straight to the HW for the week (start right away). Go back through the slides as you work through the HW for the week. The question order in the HW generally follows the order of topics in the slides (i.e., later HW questions generally correspond to later portions of the slide deck). Ignore any parts of the slide deck that aren't directly relevant to the HW (what are you hoping to get out of someone's hastily composed slide deck when the course affords no chance to apply any of that content in the HW? You can read / skim through it if you like, but that's about as much as you're going to get out of it). Just treat the HW on Bayesian methods like any other homework. No, you are not going to learn Bayesian statistics from this course (you're not even going to half learn it, in fact). But you also don't need to do that in order to get through the corresponding homework assignments. Don't panic and start asking for the title of a million different books on Bayesian methods; and really, why ask these people? Do they *seem* like folks that have any interest in - or capacity for - clear and effective communication? Their recommendations are likely going to be as useless as the lectures they put together. Do the homework, move on, research material on Bayesian methods yourself until you find something that you can actually access, and go back and learn the topic properly on your own time. That's your best bet. Aim to get a perfect score on the first 9 assignments. If you do, then - with the way the course is graded - you're done. >= 90% is an 'A', so you can hop off at that point. The topics in the final three weeks are important, but the homework assignments are not (None of them are, really. They're in general very poorly aligned to the content of the slides, and are a better tool for assessing the knowledge you walked in the door with than anything you were supposed to learn in the course itself). So read up on that stuff using a resource of your own choosing, but quit doing work for the course at that point. It's a waste of your time. Code everything in R (or Python) yourself, from scratch. You'll learn more doing it that way, and the provided code is pretty much uniformly bad / well nigh indecipherable. _________________________________________________________________________________________________________________________________________________________________ Ok - 9 weeks into the course now. None of my subsequent experiences in the course did much of anything to alter the opinions expressed above (reinforced them, if anything). I can add to the above that - after finally looking at the office hour recording for the first time in week 9 - you should definitely attend the office hour only if you're gunning for the full 'diploma mill' experience. In the OH recording that I skimmed through, I was shocked to see that the TA for the course basically just gave students in attendance the answers to the homework questions. And it's not as though this was being done at all surreptitiously. The OH recording was made available to all students via a link left in the course Piazza all semester long. So though I didn't go back through the other OH recordings for weeks prior to week 9, it's likely that many students were essentially being spoon fed the answers to the homework questions all throughout the semester. Frankly, I expected better from an 'elite' public flagship like UT Austin, but oh well. You get what you pay for, I guess.

Pros: 1. No tests or quizzes, 12 homeworks are the only grade 2. Not super time consuming 3. Cons: 1. Peer evaluation on most homework assignments 2. Undocumented provided R code 3. Frequent derivation problems on homeworks Detailed Review: This is a math course. There aren't really any lecture topics about practical applications of regression modeling. If I got an interview question about regression, this course did not really prepare me for that. That being said, it is a pretty simple course structure with only 12 homework assignments where you get 7 days to complete them. Piazza was not super helpful this time around. The TA was good but the professor would be pretty rude to people by providing non-answers and accusing them of not watching the lectures in the Piazza. There is no textbook, so if you are stuck you basically have to Piazza or Google your questions. The provided R code is undocumented and usually raises more questions than answers. I always started clean when working through the coding problems. You can get partial credit for showing your work and being on the right track on the peer evaluated homework, but some peers are more generous than others and your grade will be dependent on who you get.

Pros: 1. Teaching staff active and responsive on piazza 2. Reasonable workload Cons: 1. Some of the homework felt completely pointless 2. Lectures were a cure for insomnia 3. Did not follow any textbook, so anything that I didn't understand in class did not have a clear alternative and required looking through multiple sources to understand. 4. No context was provided for how these types of problems fit in real scenarios and how to apply these techniques.

Pros: 1. Pretty low time commitment 2. No finals or exams Cons: 1. The HW is usually peer graded, and poorly worded 2. The professor is not active on Piazza 3. The content is too vague and never really explains anything well, so I don't feel like I actually learned anything Detailed Review: The professor put as little effort as possible into this class. All but two of the HWs are peer-graded, and the questions are extremely vague, which makes it hard to know what you are supposed to put in your answers. The office hours devolved into "what slide has the answer to this question" pretty quickly, and I don't feel like anything was actually learned. Disappointing for a foundational class.

Pros: 1. Fundamental methods of the regression analysis are mostly covered 2. Workload is reasonable (except the last 2 weeks of summer session) 3. Cons: 1. Peer grading policy 2. Extra time needed to confirm the materials ( to be explained below) 3. Unclear target of this course Detailed Review: Lectures: The slides are not clear with full of typos, and often times missing necessary assumptions and details. Without providing helpful supplements to the students, this class is not suitable for beginners to dig out the materials themselves. However, for people with math/stats background, lectures omitting most of the proof is not helpful either. Not sure whom is this class designed for. Homework & Peer grading issues: The purpose of homework seems lacking purpose as well. Some required very complicated calculation (not mathematically difficult, but just complicated and time consuming). Since the lecture notes are missing lots of details, many students raised fair questions but received responses from TAs or instructors with little respect and patience. One particular TA answered without fully understanding the discussions(no matter for grading or for the questions). The reference codes provided are not helpful at all. The naming of variables are lacking comments and logics. It is hard to use. The grading policy was horrible at the beginning by just taking the very first two grades submitted. The instructor changed the policy by taking one more person's grading but the unfairness could be observed here and there. I still do not understand why they set the grading options by only good(full points)/fair(half points) and poor(no point). Since the lecture is not self-contained, I would not recommend you to take it as a review course if you don't have fresh memory of stats/math from college. I did not find this course helpful because its neither theoretical enough or practical enough. It seems like the instructor needs to work more to update his lecture.

Pros: 1. Pretty good theory course 2. Manageable workload 3. Prompt Piazza responses Cons: 1. Almost no applied examples 2. Tight scheduling - 8ish day window from material release to due dates tricky if you're juggling other responsibilities 3. Peer grading Detailed Review: So look, if you want to know the details behind why various types of regression are mathematically sound, this class class is great. If you want to know how to use various types of regression in practice, the class is pretty much useless. I'm biased because I would have preferred much, much more focus on applications. But hey, it is what it is. Given the course material, the class was ok - lectures made sense, staff was reasonably responsive (if sometimes defensive) on Piazza. Office hours fairly helpful too. Assignments are majority peer graded, which on proofs and stuff is hard (really tricky for people just learning the material to tell if an approach that's different from the answer key is valid) At some point, I stopped watching the lectures and just read the lecture notes, referring to the videos only for tricky parts/things on the homework. Didn't regret that choice. Read the supplementary material too - maybe before engaging with the lecture materials. Skip the variable selection homework - it's by far the hardest.

Pros: 1. Lots of topics were addressed that will likely be seen in future classes (maybe?) 2. Easily took less than 10 hours per week to complete lectures/HW 3. Some of the HW was purely MC on edX, so no need to worry about peer review Cons: 1. First iteration had a few bugs 2. Sample code wasn't documented well at all 3. I'm not sure that I could apply any of this knowledge outside of the academic setting 4. Not all peers are well-equipped to assess your homework even with a key 5. Grading rubric needs at least one more option to allow for minor mistakes. Detailed Review: I think for a first run of the course, it was a good run. Ultimately, I was able to complete all of the assignments with high marks, so the content must have been available somehow because I certainly didn't come into the course with the knowledge. That said, I'm not sure that I'm leaving the course knowing how to apply it outside of the confines of academic coursework. Perhaps that's not the purpose of this course and another course in the program will serve that purpose. Interaction with Dr. Walker on Piazza was generally brief. He was active in spurts, but it took time to understand his responses. I think he's fairly knowledgeable on this content but there is a disconnect (most likely because of the modality) between what is conveyed and what is understood. One could argue that's reasonable in a graduate level course. Coming from DSC 381 last semester, there does seem to be a difference in pedagogical practices, but the reduced time of the summer semester might have contributed to this.

Pros: 1. Workload is okey. Cons: 1. Homework was lacking purpose. 2. Provided code is bad. 3. Lectures had a lot of typos. Detailed Review: Generally, I felt like the assignments were lacking purpose. Mostly, I learnt things that I'm unsure of how to apply. "Monkey see, monkey do". 50%-75% of homeworks are derivations and integrations to get the same equations presented in the lectures. Homework didn't have interesting purposes nor datasets. Datasets were very obscure, only consisting of "y" and "x" (usually one x column), I found this very boring and hard to translate to real world applications. The provided code is completely useless. All of the variables' names were non self explanatory. You had to guess what the variables were refering to as they all had obscure notations such as "x", "xx", "xxx", "xhh", "yhh". Long story short, you had to make your own code because the provided one was very hard to understand. Lecture slides had a lot of typos, specially on equations (missing square term, missing variable, inconsistent notations, etc). A lot of things were not explained in lectures and where described as "straightforward" when they weren't trivial. Lectures were also missing "real world" examples and some lectures didn't even include an example. This course needs a lot of improvement. I was expecting for an applicated class, with lot of real world examples or interesting free datasets available online.

Pros: 1. General regression topics covered are applicable to a wide range of problems 2. HW assignments and lectures are of reasonable length Cons: 1. Class content is not put into real world context 2. Order of presented material doesn't always flow well week to week 3. The sample code provided in the course is garbage 4. The notes were riddled with errors (first iteration of the course - hopefully non issue in the future) Detailed Review: This course really opened my eyes to the breadth of topics covered in regression and how complex (but useful) this knowledge could be. The professor does a good job of covering a wide range in relatively short amount of time but there are lots of "holes" left to be filled knowledge wise. A lot of "expertise" building though will require extracurricular activity and the course did not provide much in terms of finding resourcing for that. This course will also be a struggle for those who don't have basic programming skills (in either R or python) already developed. The sample codes provided by the lectures lack annotation or clear variable naming and can be hard to follow. This course was very much a "math" course. It does not touch on real world example problems, nor how to select/deploy models in working environments. The course content was very "backloaded" with the 3 last weeks containing the hardest (and probably most useful) content. It was also condensed roughly into 2 weeks during the summer semester. Homework wise - it is a mix of multiple choice and peer grading. This semester had no exams (all HW-based grades). Peer grading could be harsh depending on the competency of your grader. Difficulty wise - this course falls between Probability and Machine-Learning. Piazza got very difficult to navigate as weeks started to overlap. Instructor replies were sometimes useful, and sometimes not.