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

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

Loading...
  1. Dustin

    Dustin Well-Known Member

    If you used R, you'll probably be familiar with most of the datasets we're using in this course because they come with the ggplot/tidyverse libraries: nycflights13, msleep, pizza (this was a CSV so I'm not actually sure if it comes with), fastfood.
     
    JoshD likes this.
  2. JoshD

    JoshD Well-Known Member

    Very familiar with nycflights13 but not the others. Other than nycflights13, we used COVID data from North Carolina, election data, etc.
     
  3. JoshD

    JoshD Well-Known Member

    Dustin, these are the topics we are covering in my Data Analytics and Applications course this term.

    663D24FB-C8B7-4C8F-BBD1-60602BCFB6DC.png
     
    Dustin likes this.
  4. Dustin

    Dustin Well-Known Member

    Neat! This material is covered at Eastern in the Fundamentals of Machine Learning course I'll take sometime after this one.
     
    JoshD likes this.
  5. kcshocker

    kcshocker New Member

    Hi Dustin, thanks for all of your updates about this program. About how much time have you needed to devote to this program each week? I'm trying to figure out if I have the ability to complete course work since I have a full-time job and a family. Thanks!
     
  6. Dustin

    Dustin Well-Known Member

    I'd say I put in 30 minutes a day during the week and then 2-4 hours on the weekends. One course is very doable for someone who works full time. 2 courses would be challenging, especially the later ones.

    For 650, the course I'm in now, the first set of coding assignments took me 8 hours. I estimate the second set will also take similar.

    The actual content is 7 hours of video plus study time for the 4 exams.
     
    kcshocker likes this.
  7. kcshocker

    kcshocker New Member

    Ok thanks. That is good to know. I'm considering taking 2 classes at a time for the first two 7-week sessions and then dropping down to 1 class at a time. I know basic Python but my math is rusty. Bone head idea or doable (based on what you've seen so far)?
     
  8. Dustin

    Dustin Well-Known Member

    Definitely doable for the first 2. They're exam-based and theory-based. The 3rd course and on involve exams, code assignments and final projects so I wouldn't do 2 at a time myself after the first 2, but those first ones are foundational courses for people with no exposure to Python or stats so you should be fine.
     
  9. kcshocker

    kcshocker New Member

    Thanks for this advice!

    I was accepted into the program. But when I spoke with the non-academic advisor, she informed me that I should have a solid foundation in statistics prior to entering the program and that advanced math is a "large portion of the program". This surprised me as it isn't mentioned anywhere on the website. I mentioned that I was rusty, and she said "you will definitely need to brush up on statistics first." This seems to contradict the "meet you where you are" aspect they so heavily promote (and was the reason I applied).

    Based on your experience so far, would you agree with this advisor?
     
  10. Dustin

    Dustin Well-Known Member

    That is really surprising to me. No, I don't think you need a lot of stats knowledge. This is coming from someone who never did any kind of hypothesis testing or anything before I got into this program. The math is very applied. You'll learn lots of stats, but it's always in the context of why you need to know this and only enough to know how to use it. 550 (the first stats course) starts with mean, median and mode and goes through probability, probability distributions, hypothesis testing, t-tests, f-tests, chi-square, ANOVA, etc.

    I'd never done a t-test before I did that course. Now I know how to do one and interpret it, though.

    650, the course I'm in now is more advanced but it's still approaching the topic very gently. The sheer amount of material that appears to be covered is a lot, but it only goes into what you need to know about those topics. So for example, in 650 we learn about Spearman's correlation. Why we need to use it, how we calculate it and what it means. But we only saw the formula, we never need to calculate it ourselves (because R will do that for us.)

    Same thing with point biserial and biserial correlations. We learned when and why we need to use them, but not the theory behind them except as it applies to data science. We do learn about linear algebra and some calculus in the machine learning courses (I've been told), but nothing that you can't pick up yourself.

    Edit: I should note that the first time I ever learned about Pearson's r and Spearman's rho and point biserial correlations and chi square and ANOVA and generalized linear models and parametric data and the Akaike information criterion (AIC) and 90% of the stats stuff in this program were in these courses.

    Most of my fellow classmates are not math or stats people and most people are not struggling.
     
    Last edited: May 28, 2021
  11. kcshocker

    kcshocker New Member

    Thank you again for such helpful information!

    I emailed the program director to ask his take on it. He emailed me back within minutes (awesome!) and said something very similar to what you are saying. He reassured me that there is no expectation of a math background for incoming students and that the first class "starts from zero". The overall focus is on applied skills rather than being bogged down by math, even for the machine learning courses. He wrote that many people with less math in their backgrounds have succeeded in this program. This seems to be the program I had hoped for.

    I feel much better now.

    Again, thank you for all the info you have posted on here. I am eager to hear how things go in the future.
     
  12. JoshD

    JoshD Well-Known Member

    I am not in this program but I can tell you that many programs begin at ground zero. Literally, learning how to create a vector and data frame is a part of introductory courses which, as you will learn, is very simple. However, they build upon that and it can get deep. There is so much to know about data science that there is an expectation that you practice and learn outside of that course as well.
     
  13. Dustin

    Dustin Well-Known Member

    Completed the 8 modules for this Data Analytics with R course:
    • Intro
    • Visualization with ggplot
    • More R
    • Model-Building
    • Assumptions
    • Correlation
    • Linear Regression
    • Logistic Regression
    I also completed 4 basic CodeGrade assignments in R and 2 of the 4 stats CodeGrade assignments. All that's left is the 2 CodeGrade assignments and the final project. Currently have a 94% in the course.
     
  14. Unistic

    Unistic New Member

    So you would consider this course relative easy? Even a non math/IT major could do it? Are any of your classmates taking two courses at a time? I plan to use my GI Bill (just applied for the degree today) but I only have 16 months remaining and I don't want to pay out of pocket which means I'll have to double up a couple times to complete it. I don't want to take two at a time but taking two will ensure I don't come out of pocket.
    Also you didn't mention having to write papers but just eval type assigments/quizzes. You also mentioned now that they have proctored test which I would guess they didn't have it before.Are you able to repeat the proctored test as well until you achieve the wanted score? Lastly you mention GA status or something, I've read on Reddit that its some type of advisor position for fellow students in the previous course and that you receive some kind of course payment skip or something can you please explain that? Thank you so much and for all the information on this degree you've provided hoping I get in!
     
  15. Dustin

    Dustin Well-Known Member

    It's challenging in that you're learning new skills you haven't had before and that takes hard work but it's not needlessly difficult and the math is wrapped in plenty of examples of how you use it for data science. I'm not a math person and I'm not struggling (yet, at least!) You learn the skills you need and nothing you don't. Yes, lots of people take 2 courses at a time but they have only work and school. For those with more demands on their time, 1 course is probably better. You can double up on at least 520/550, and 575/600. 660, 670, and 680 are very hard to double up on.

    There are no papers, but 690, the Ethical Data Science course that makes up half of the capstone, does have discussion posts as assignments.

    Proctoring uses Respondus. It's virtual proctoring so you record a video around your desk and then the system records video of you taking the test. If it detects someone on the screen or audio, that portion of the video is flagged for the Professor who then checks to make sure you weren't cheating. They're pretty relaxed about it. Yes, you can repeat the tests as much as you need. I've had tests with 10 repeats to get the highest grade.

    Graduate Assistants are students who completed that course already with a high grade, who answer discussion board posts in the lower courses and help out with grading in the upper courses. GAs are selected at the end of each term. You can only GA one course at a time, but for each term that you GA, you will be given a rebate for one course. So if you're only taking one course in a term and you're also a GA, that term will be effectively free. I'm currently a GA for 520, the first course in the program.

    Hope that helps!
     
  16. Unistic

    Unistic New Member

    Thanks alot! I'll keep those course in mind when looking to double up! I'm retired military so I have the time to devote to studying but i do have a hard time procrastinating which I"m hoping to get over. No papers and more project/exam base is what I would prefer as well. Thanks again!
     
  17. Courcelles

    Courcelles Active Member

    How’s the religious content there? Noticeable, forced a la Liberty, or completely avoidable?
     
  18. Dustin

    Dustin Well-Known Member

    Absent in the course content. There is no religious content in the program at all, but they are a Christian university and the Professors all agree to a faith statement to teach there. As part of the application, you'll be asked for your religious affiliation, but Agnostic was on the list (that's what I chose), and I wouldn't be surprised if Atheist was on there too. Religious affiliation or faith community does not affect your ability to attend, according to the Admissions staff I spoke to.

    As an enrolled student, you'll get an invitation to the Daily Prayer on Zoom to your Eastern email, but of course attendance is not mandatory and I've never been. The President occasionally sends out requests for reflection or prayer in campus-wide emails.

    Otherwise, the only references to religion in the course content so far that I've seen in the first 3 courses are references or examples of Professor Longo's actual research on religiosity, and it was always used in in an illustrative way where it was a good example of the concept we were discussing.
     
  19. Dustin

    Dustin Well-Known Member

    CodeGrade assignments complete! Those last few were a doozy. Often I was adjusting output to conform to CodeGrade (especially around rounding), or I was using standardized instead of unstandardized (or vice versa) coefficients or otherwise producing the wrong output. I think I've got a decent handle on how to generate linear and logistic regressions now, but interpreting them will take some more work.

    With only the final assignment remaining, I have a 95.8%. In order to finish the course with a 93% or higher, I'll need to either redo one/both of the exams I did the worst on or get a 68% or higher on the final. Ideally both.
     
  20. Unistic

    Unistic New Member

    Re-asking previous questions, the final can not be repeated. But the way you previous stated about the requirements of making an 80% on each quiz before progressing to the next and the fact that you can retake the previous quizzes until you're happy with your score and with the weighted grades of the quizzes/projects the final example is less weighed. Meaning if you're able to retake everything else until you make a 100 if you see fit then when it comes to the final example even if you make a 30% (meaning everything else is 100% scores) you still pass the course overall. Am I correct in that thinking?
     

Share This Page