Why do data scientists need to learn R?

I hated R the first time I saw it. It seemed to me it wasn’t as community supported as python, I struggled to learn the language by myself. But I’m taking my second graduate course that uses R and teacher has taken the time in the schedule to explain R.

In the very beginning I didn’t understand why this language was so important to learn for a data scientist, a language I had barely heard of. But only when I got to use it was when I realized why. 

It’s so simple, so elegant, so straightforward. You can do so much with such a few lines of code, sometimes I don’t even miss Python, and that’s a something that I cannot say of most languages.

I haven’t gotten a chance to work with Scala yet, that’s one of my goals for 2018, I hope it’ll be as enjoyable as what Phyton and R have become for me. It’s being a whole journey to become a data scientist, loving every minute of it, and loving to share it with you guys. See ya next time.

Why do data scientists need to learn R?

What was it like to take 1 OMSCS + 1 OMSA micromaster course in summer 2017 @GeorgiaTech?

omscs-vs-omsa-2

Hi again guys. Long time no see.

Sorry I haven’t written anything for a while. The reason for this is I took 2 graduate-level courses in summer (you know, with a family and a full time job as usual).

How was it? Let me put it this way: It was hell on Earth. I even made the crazy decision to go in a vacation trip (bad idea). I had to do a coding project during my time in 2 different countries, 2 different hotels. I was a tourist by day, student at night.

As you may already know, OMSA is the Online Master of Science in Analytics, a very interesting master from Georgia Tech that caught my eye from the very first time I heard about it. The OMSA first semester won’t start until Fall 2017 really (by the end of this month), but I found out that unlike OMSCS there is a MicroMaster from edX called Analytics: Essential Tools and Methods, which includes 3 basic courses from OMSA that can be applied for credits once accepted in OMSA.

The MicroMaster course I took is called Introduction to Analytics Modeling (ISYE6501x). I took the course knowing almost nothing about it, only a short description from the edX website. I looked for the syllabus online but had no luck.

On my first week of summer “semester” I was really anxious to grab the schedule from both courses (Reinforcement Learning from OMSCS  and IAM from GT’s MicroMaster at edX). That was when I found out about the workload from each course:

  • Reinforcement Learning (OMSCS): 6 homeworks, 3 projects and 1 exam
  • Introduction to Analytics Modeling (OMSA): 10 homeworks, 1 project and 3 exams
  • Total: 16 homeworks, 4 projects and 4 exams in 10 weeks for me

That included a 2 week vacation trip I had already agreed to. That was the precise moment I got like nuts haha.

Besides that, we had to grade our fellow students in IAM class, 3 grades per assignment which would be averaged in order to get our definite grades.

OMSA vs OMSCS platform

If you are an OMSCSer yourself, you’ll know that we at OMSCS use different independent systems for every different “necessity”, which I show below:

  1. OSCAR.- To enroll in courses
  2. Piazza.- For course discussions
  3. T-square.- for assignment tracking
  4. Udacity.- to watch Lectures and small quizzes

and so on…

Sometimes we even need different passwords for each application, like Piazza and OSCAR. But in OMSA, you use edX, which has everything integrated into it. If you are a software engineer you know the tradeoff that integration means: user-friendlier in the edges, but less features and less usability while you get deeper into a component.

I think it will be way easier if you let me show it to you in pictures:

This is the main menu on the top of edX courses, as you can see there are 4 main menu entries, which I’ve taken the liberty to show what its replacement would be in the OMSCS world.

edx-course-menu

Course

If you click on “Course”, something like the picture below will show up. From my point of view it’s way cooler than Udacity, as we don’t only have the lectures in front of us, but they are divided in weeks and each week has its homework and homework solutions afterwards, like a combination of Udacity + T-Square shown in a chronological order, very sleek indeed. Besides Udacity’s usability (human-computer interaction) is not the best, at least for my taste, it can be annoying sometimes.

content

Good: The videos in edX are very well polished, from my point of view the video production is a little bit more elegant than what I’ve seen from OMSCS so far in the 5 courses I’ve taken.

Take yourself a look at 2 different videos from OMSCS and OMSA, you be the judge:

Bad: There is no “unlimited resubmission” feature available for assignments. That’s a major feature where t-square is just superior to edX.

Discussion

The ugliest part of edX is maybe the Discussion Forum, let me show it to you, compared to piazza it’s just well, how can I say this without being hurtful? Let’s say it’s not piazza, not even close.

discussion

Good: It’s a usable forum

Bad: It’s not piazza

Wiki

Actually we didn’t use it in this first course, so I cannot tell you anything about it.

Progress

It’s a graphical view of how your overall grade is doing over 100% total in time, it uses bars to show your scores on every assignment and exams, and your final score. This would be like T-Square’s Gradebook, but adding the visuals.

Course Execution

The part I had the most problems with is that based on the short initial description of the course before I got enrolled, knowing R language was no pre-requisite and I didn’t know anything about it really, but on first week of the course we were given our first homework which included to deliver working R code due for the next week. People complained on that one (including me), as we didn’t expect it to be like that, and we had to struggle a lot to learn the language on the run and make it through the datelines.

Conclusions

I learned a ton. The first lesson learned was that I should never again enroll in 2 graduate-level courses in summer while working full time or going on vacation lol. I passed both courses, but it was a real nightmare, there was even a time I thought I would have to dump one of them in order to pass the other one.

Second thing I learned was that OMSA is a very promising masters and I’m really insterested on that, I’m looking forward to take another micromaster course. I found IAM to be a very usable course. From my point of view it’s more usable in a daily basis than courses I’ve taken from OMSCS so far.

Additionally, I learned I should not underestimate a micromaster course workload. IAM wasn’t hard but it took quite a lot of time for me to pass the course (watching lectures, doing assignments and 3 exams, with final being cumulative)

In sum, it was quite an experience, I wish I could have gotten more time for every class so I could have learned a little more in detail, but as I had so many assignments, I had to do a lot of prioritization, otherwise I wouldn’t have made it through.

Hope you learned something from my experience, I know I did. See you next time.

Update (Aug 10, 2017): Just got the edX certificate, and this is how it looks like:

 

 

 

 

 

What was it like to take 1 OMSCS + 1 OMSA micromaster course in summer 2017 @GeorgiaTech?