MS in Data Science: self-paced, 10 months, under $10k

Discussion in 'IT and Computer-Related Degrees' started by Seylan, Jun 5, 2020.

  1. Dustin

    Dustin Well-Known Member

    The first 3 courses require no purchased textbook but provide links to free resources. The first course is made up of 6 modules:
    • Intro to Data Science
    • IPython
    • NumPy
    • Pandas
    • Matplotlib + Seaborn
    They use a series of professor-recorded videos with embedded quizzes and an exam at the end of each module. It's not proctored (yet) but the syllabus provides the option for them to implement proctoring. You must score 80% or higher on the previous quiz(zes) to be eligible to write the next one, and you can repeat them until you get 100. There are assignments required at different points in the course, and you only get one shot at those.

    The video lectures are short (10-20 minutes each) and total about 2 hours per module, I haven't gotten to a quiz yet. You have until the end of the semester to submit (so in this case Feb 28). The content included here looks like most of my DataCamp Data Analyst with Python course I took (with the exception of basic stats which is covered in the next course.) In the first module is an introduction to Google Colab which I'd never heard of, but allows you to offload your processing to Google servers, and otherwise is identical to Jupyter notebooks if you're at all familiar with those.
  2. sube

    sube Member

    Thanks, that's interesting. It sounds similar to taking a course on Coursera or one of the other MOOCs since it's self-paced and exams and quizzes are embedded. Do you have to do any projects with other students or post in discussion boards? I've never heard of Google Colab either but I will check it out. Is there any interaction with the professor or do you only contact him/her if you have a question?
  3. Dustin

    Dustin Well-Known Member

    Discussion boards are pervasive. There is a parent MSDS "course" with a discussion board for the whole program. Each module has a discussion board (e.g. NumPy discussion board for questions about that module.) Each course has an overall course discussion board. And there is an Eastern University discord, unsanctioned but promoted by the professors (and run by students.) Discussion posts are optional but if you provide 10 substantive replies to content-related questions, you're eligible for a 2.5% boost to your letter grade.

    There are 2 capstone courses, an Ethics in Data Science project where you develop a proposal and learn about how to do ethical data science projects (this is the second-to-last course) and a capstone where you develop your proposal into an actual project. You can collaborate and study with each other throughout the program but don't have to. The capstone can be done in groups but doesn't have to be.

    There are 2 Graduate Assistants in each course who got an A in the course previously and have strong communication skills. In exchange for being a GA in a course, they get a tuition rebate for another course. They answer questions on the discussion boards, and the Professors make themselves available via email generally. The course materials are all recorded lectures by the professor, so in that way there is some interaction, but generally you contact them if you have questions.

    All in all, they're very flexible.
    nomaduser likes this.
  4. Dustin

    Dustin Well-Known Member

    Here's what the lectures look like. They're all recorded, and they use Jupyter to show you the code, talk about it and then show you immediately how the output changes. I really like the format.

    Attached Files:

    SteveFoerster and nomaduser like this.
  5. JoshD

    JoshD Well-Known Member

    Jupyter is not bad but I prefer Spyder. Probably because it very much resembles RStudio in appearance and use. Lol
    SteveFoerster likes this.
  6. sube

    sube Member

    I saw your other posts about Quantic. You're a busy guy! I had never heard of that one, so I'm taking the free courses there now. I like their learning model and format. Going through the class is actually fun! Anyway, thanks for posting all this info on Eastern. I'm approaching the end of my career so getting a master's is not something I really need at this point and I'm not sure I can justify spending the money, but I will think about it more before I make a final decision. How would you rate the rigor so far or is it too soon to say? Are they starting you off with the basics?
  7. Dustin

    Dustin Well-Known Member

    It does starts with the basics, but it feels like there's a bit of a learning curve for people not familiar with programming. I know basic Python, so it's mostly about getting re-familiarized. Too early to say the rigor. It definitely hits all of the topics for someone with no background in data science: Python, NumPy, matplotlib, seaborn, IPython, how to use Jupyter, R, descriptive statistics, probability, multiple and logistic regression, Qlik, Tableau, other regression and classification techniques, SQL, database management and cleaning data, k-nearest neighbor, SVM, decision trees, PCA, neural networks, deep learning, and two capstone courses.

    But with that many topics covered, I think the bigger issue is that it's hard to get indepth in any of those topics in 7 week courses.
  8. Dustin

    Dustin Well-Known Member

    I completed my first "exam" (to differentiate it from the non-credit quizzes.) It's 50 questions, auto-graded and scored immediately. A mix of fill in the blank and multiple choice.

    Very tough. Very little opportunity to guess. You have to really know the material.

    I got a 74 on my first try and an 84 on my second. Someone on the Discord mentioned it took them 10 tries to get 80%+.

    At least one student was actively dropping out. He was in the first conort and doing 2 courses at a time so he was 60% through (January is the 3rd semester they've been offering the program) but decided to switch.

    He had considered WGU, GA Tech, and the other usual suspects but I'm not sure where he decided to move, but he was leaving data science.
  9. sube

    sube Member

    Hmmm, that's interesting. It doesn't seem like this is really for an absolute beginner (as I am) so I'm going to pass. I really only need business analytics, not data science, so I'm going to look into that, but I'm intrigued by the Quantic program. I'll post something in that thread since this one should just be for data science.
  10. Dustin

    Dustin Well-Known Member

    I do hope you find a good program for you. I was worried I couldn't find a program that didn't require Calculus and Linear Algebra. I know that those are important to understand what's happening under the hood but data scientists are often using libraries and need to understand how to use a library but not necessarily how to build. It's practitioner vs researcher.
  11. Dustin

    Dustin Well-Known Member

    I finished my first course. My grade is a 92.3% currently. I understand somewhere between 91 and 94 is a 4.0 so I sent off an email to find out what the exact percentages are, because the school doesn't put that in that catalogue, just the letter grades (e.g. an A+ = 4.0, A = 4.0, A- = 3.7, B+ = 3.3, etc)

    Edit: That was quick! They emailed me back and pointed me in the direction of the syllabus. A 93 or higher is a 4.0, 90-92 is a 3.7.

    The exams got easier as I went. I have an 83.8 in the first exam, a 91.3 in the second, a 94 in the third, and a 100 in the fourth (and you can repeat them until you get the desired grade, or until the course ends. Whichever comes first.)
    Last edited: Jan 24, 2021
  12. Dustin

    Dustin Well-Known Member

    Finished my first course. I got a 93% so I finished with a 4.0. I also learned I was selected to be a Graduate Assistant in the next course - pretty cool. In exchange for being a GA you get a tuition rebate, so my next course which starts Monday will be free.
    Graves, sks972, JoshD and 2 others like this.
  13. Dustin

    Dustin Well-Known Member

    Update: Completed 3/5 modules for this course so far. Right now I'm averaging right around the 80% mark on my exams so I'll have lots of time to revisit these exams to get my grades up (the course ends April 25). This is a basic stats/probability course in R. The exams focus primarily on the stats but there are labs we can do to work in R and they'll give us the answers once we've made a good-faith effort to do it in R ourselves.

    Content includes:
    • Basic probability and probability distributions
    • Measures of central tendency and variability
    • Hypothesis testing and statistical inference
    • Chi-squared tests of good fit/independence
    • t-test
    • ANOVA
    Graves and Maniac Craniac like this.
  14. Graves

    Graves Member

    Thank you for the updates! I'm strongly considering applying for this program in 2022.
  15. Dustin

    Dustin Well-Known Member

    I'm liking it so far. I ended up finishing this course with a 93. It took me 14 exam attempts to average a 93 across the 4 modules. In 550 (the second course, the basic stats one I just finished) it took me 32 tries across the 5 modules (between 3 and 10 attempts per each module), mostly because it took 10 attempts to get the second module, focused on probability distributions down.

    The first course is a general intro to DS and the theory of data science (with some basic Python.) The second course is general intro to statistics and probability with some basic R. The courses that come after are applied R and applied Python.

    The course doesn't officially finish until April 25 and there's a break in between the terms so I won't be starting my next course until May.
  16. sanantone

    sanantone Well-Known Member

    It's because U.S. News factors in SAT/ACT scores. Texas A&M has to follow the top 10% rule. If you graduate in the top 10% of your class, you're automatically granted admission to any public university in Texas. So, some Texas A&M and UT admits didn't take the SAT or ACT, or they did take one of those tests and scored low. Public universities in Texas don't have to follow this rule for graduate programs, so they can select who they want.

    Interestingly, this has led to some wealthy students choosing to go to low-performing schools so that they can easily graduate in the top 10 or 6 percent.

    Note: UT has been granted special permission to only admit the top 6%. The Texas legislature wanted to make the school more competitive nationally. You can't admit too many out-of-state students when so much of your student body consists of Texans.
  17. Dustin

    Dustin Well-Known Member

    I completed the second course in the sequence, DTSC-550 with above a 93% so my 4.0 GPA remains intact. Next course (for me) is DTSC-650, which is applied statistics using R. I was again selected as a Graduate Assistant for the next term. In this most recent term they started using proctoring software called Respondus Lockdown, which aims to prevent cheating by having you videod through your webcam while you complete your exams. It's much less invasive than ProctorU or another kind of live proctoring.
    sks972 likes this.
  18. Dustin

    Dustin Well-Known Member

    Started the next course. It's been a while since I made an update (thanks Josh for the reminder!) This course is Data Analytics with R. It's made up of 8 modules:

    • Module 1: Course Intro
    • Module 2: Visualization with ggplot2
    • Module 3: More R
    • Module 4: Model Building
    • Module 5: Assumptions
    • Module 6: Correlation
    • Module 7: Linear Regression
    • Module 8: Logistic Regression
    There's a test for each module (except the intro one), and 4 auto-graded coding assignments. Additionally, there is a "one shot" assignment that, unlike the others in the course, cannot be repeated. You only get one try.

    Completed the first two modules. Module 3 is supposed to have the most difficult exam in the course.
    SteveFoerster and JoshD like this.
  19. Dustin

    Dustin Well-Known Member

    Productive weekend!

    I finished Modules 3, 4 and 5 with some late nights. They either updated the exam for Module 3 or I had wrong information because I was told it was the hardest in the program but I had no issues with it at all. I also got a set of coding assignments done in R. There are 5 assignments. Each one starts with a dataset and then asks you to do standard questions, so for a dataset with information on flights in NYC it would say something like, "Calculate the mean distance for flights taking off from JFK and ATL airports" or, "What were the longest and shortest delays?", things like that.

    5 different datasets, each with 5 questions each. Because I haven't used R before, the first assignment took me like 8 hours to do. The second took me 2 hours. The third, fourth and fifth took me 2 hours. Now I feel like an old hand at it. There is another set of these assignments coming up after Module 8, that focus on statistics, and then a final project worth 10% of the grade that isn't graded until the course is over. The final project is a real dataset from the CDC and we'll be doing summary statistics, regressions, correlations, model selection, etc. Pretty neat.

    Started learning about modeling data, assumptions of normality, the Shapiro-Wilk test of normality, homogeneity of variance (and Levene's test for homogeneity) and how to implement those in R.
    Now I'm on Module 6, correlation.

    Currently have a 94 in the course. Aiming to keep that 93 so my 4.0 will stay intact.
    SteveFoerster likes this.
  20. JoshD

    JoshD Well-Known Member

    The Flights data is a very common assignment for R. We utilized the same dataset in my Programming for Data Analytics and Visualization course.
    Dustin likes this.

Share This Page