cs373 "effective" students (by surveying the forum...)


AFAIK no "official" number of enrolls in cs373 have been reported. Also, the hottest questions in the forum seldom raise more than 1K to 2K views. So, I went to check the information available to figure out what that number could be. Below are some findings.

Notice that there can be people doing the course without being registered in the forum, but this is awkward, given the number of "incidents" related with grading bugs, etc... which are discussed here.

There are 127 pages of users. Each holds 35 of them but the last, which has only 7 right now. So
there are 4417 registered users in this forum. (NOTE: cs101 has 7492 users in their forum.)

Regarding activity (karma points), only in user's page 73 (running down from 127) the karma starts raising above 1. This gives us approximately 2555 users with karma > 1, that is, only about 2500 have more than an infinitesimal activity in the forum.

A karma=100 is in users's page 17, which gives about 600 users with a sensible activity (of course the limit=100 is empirical, use the one you want...)

In the end, how many students will have submitted all of the HWs till now? Any guess? Does the "grading policy" of Udacity which now accepts an "exam-only grade", discarding homeworks (if worst than the exam grade), conveys the idea that it's enough to do the final exam to get the "diploma"?

This is just "smalltalk", and a silly question for those who already submitted HW5 :-) In the end, the number of active students in "Programming a Robotic Car" is perhaps a hundred times larger than the number of those attending any other similar course in the world...

UPDATE: 4695 users registered in the cs373 forum by 06 of April 2012. Probably the "saturation point" has been reached, since the final exam "closes the course" in 3 days.

asked 25 Mar '12, 21:57

Jose's gravatar image


accept rate: 0%

edited 05 Apr '12, 22:47


You don't have to sign up to the forums to read them though. For the AI course, I didn't even sign up to aiqus until after the course was over as I had nothing to say and I don't suppose I was by any stretch of the imagination the only one.

(25 Mar '12, 22:43)

fnenu-1 ♦♦

fnenu-1's gravatar image

Now you are not acting the same way, @fnenu :-)

More seriously, the numbers I've shown are those available from the forum site: perhaps several geniuses in Probabilistic Robotics don't even bother to register in the forum, but I think those are small numbers.

Regarding the AI course (which I didn't attend) the number of registrations were 160000 and I think 20K or 30K have succeeded. I just found a site which conveys some information about the grading methodology pursued by Audacity, according to Sebastian's own words, and it was tailored after they (Sebastian and Peter Norvig) noticed the failure of the grading system being used in the first stages (weeks) of AI.

(25 Mar '12, 23:39)


Jose's gravatar image

You failed to tell us which is your measurement gaussian error. Without it, I can not compute a prediction.

(06 Apr '12, 00:37)


Micro-BIOS's gravatar image

By way of comparison, 20,000+ (30,000+ ?) finished the AI course, but much less than 1000 of them were active on the forum aiqus.com at the end of the course.

Hence I would suggest that there is definitely a multiplier of forum users to course users and that multiplier is is of order 10-100.

(06 Apr '12, 09:11)


SimonS-1's gravatar image

As far as I'm concerned, I only signed up to the forums when I had a question that was not answered. I have a 50+ hr a week job, a long commute, and also taking a brick & mortar course (in addition to a slight overlap with the algorithms course) on top of other things I have to do. By the time I can actually get to the homework, most of the questions have already been answered. And, I did finish all homeworks and exams, on time. I didn't have that many questions overall outside of grading issues.

With that said, I am envious of those that have a lot of time to review the forums thoroughly, help others, answer questions, and discuss more profound implications of this material. Unfortunately though, there's no time. Thanks to all who did put in that time!

My point is, you're making a lot of assumptions here, but I would be interested in results of a survey about how people interact with the class.

(06 Apr '12, 11:17)

Douglas Siev...

Douglas%20Sievers-1's gravatar image

5 Answers:

That has been a burning question for me, ever since I signed up. When I tell my friends about Udacity, and how the first AI class had 140K signup, with 40K completing, those numbers just blow them away, and I wish I had some similar numbers to relate when I talk about the current course I am in.

I think knowing how many students are "enrolled" / "active" / "completing homeworks" / "took the exam" is a useful ongoing metric to know. While it's probably not directly tied to the education we are receiving, it is one of the things that you can expect in a traditional classroom. I think the combination of how the units are designed and presented, along with our fantastic community on the forums, provides everything that we get from a "real" classroom, in a superior fashion. The only thing that I find I really miss is that sense of how many people we are with in this together.

It sure would be nice if we didn't have to speculate and extrapolate to come up with estimates:}


answered 25 Mar '12, 22:44

billi_rush-1's gravatar image


Well, in the end the figures (# of enrolled students) don't matter, only the knowledge earned by completing the course.

My personal view is that self-study is much more inefficient than enrolling in these courses, even if in the end you could get the same knowledge, because here there are deadlines (in my case they add to those in a regular work...) and the feeling of that peak of adrenalynn just before HW results makes me going back in time to my university days :-)

(25 Mar '12, 23:52)


Jose's gravatar image

Somewhere I read that over 90,000 signed up for CS101 and CS373. Based on past online courses (intro AI, machine learning, etc) about 15% will complete the homework and exam so we've probably got somewhat less than 15,000 active classmates. I'll let others speculate on the distribution of these students across CS101 and CS373.


answered 25 Mar '12, 22:14

MCN's gravatar image


It could be... there is perhaps some intersection (many people in both courses), but its hard to me to believe that the remaining 10000's of people (subtracting those registered in the forum from the 90K) are doing the HWs. They can see the forum, I think, but not to play there...

(25 Mar '12, 23:45)


Jose's gravatar image

we were talking about this in another post (http://www.udacity-forums.com/cs373/questions/32254/how-many-students-completed-the-course-and-any-chance-of-one-last-office-hours) and so far the number of views that final progrmming hw video got is 377. Also consider that yotube takes some time to update this number and there is still time to finish the exam http://www.youtube.com/watch?feature=player_embedded&v=C-3UND3ylX4


answered 05 Apr '12, 23:13

Fernando%20Saravia%20Rajal's gravatar image

Fernando Sar...

So your point is that there are about 377 students enrolled here?

(07 Apr '12, 00:18)


Jose's gravatar image

Well i thought so untill some other people said that youtube doesnt count embedded video views

(07 Apr '12, 00:25)

Fernando Sar...

Fernando%20Saravia%20Rajal's gravatar image

Dave addressed the issue of releasing the number of students attending each class in an office-hours video of CS-101. They do not want to release them because they do not think that it is a productive metric. It is a metric that can be manipulated easily and is not measuring the "right" thing. You can imagine comments of the sort of "this course (or this professor) is "better" because it got more students". They might have a point.. and in spite of that, my guess is that it is likely that eventually the numbers will leak into the media. It is indeed a useful value to attract students but that can be done by other means. For the people in the course, those who saw its value from the beginning, it is a useless metric: most of us would have taken this course regardless of whether 100 or 100k people had signed for it.


answered 06 Apr '12, 09:24

Goldsong's gravatar image


edited 06 Apr '12, 09:25

I somewhat agree with your POV. I thought the # of students was an interesting question, as would be the # of students per country, etc, and it was the "inquisitive mind" the reason behind these trivially-calculated figures I did.

I'm also doing this course because I thought it was very interesting - a point I've by now confirmed, since I finished it. I'm also involved in courses outside Udacity; I already un-enrolled one of them, and I'm following another. It's interesting to see the various approaches to online teaching and grading, and how they succeed and how they fail. I have a regular job, probably as many of others around, so in general I'll have much more experience than "youngsters" in tackling the HW and final, but much less time.

Nowadays you often have to overcome technical interviews and exams for applying to many jobs (true in IT and CS); it's not enough to have a certificate from a top notch University or Institute: you have to prove your skills. So, if you manage to getting educated online, in some years you can go through a "cheap" but good-quality education. However, you are not seeing basic Math and Physics courses being taught online right now, and those are basic education for most Science, Eng. and Tech. courses. So you cannot start your education. And there's the lab issue: how can you get good lab skills (physics, electronics, mechanics,...) online? How will you learn to debug circuits, e.g. to spot short-circuits through the smell in electronics labs, for instance, only with online courses?

(07 Apr '12, 00:13)


Jose's gravatar image


All this information should be made public, instead of having to resort to more or less informed second guesses.

These courses are excellent at the technical level (and I'm grateful for that), but the non-technical aspects, that is, the values, principles and attitudes which guide the organizers, leave much to be desired. They get an A+ at the technical level, but an F at the non-technical aspects. There's much to be improved, particularly regarding transparency and fairness, for these platforms to really be the basis for 21 century education. They are actually doing a bad job at it.

While they are evaluating us, they are also being evaluated. And sorry, they fail.

(Down vote as much as you want/can. It'll show up at the end, so it'll be easy to find. And it'll also show up in other places, you can be sure about it.)


answered 06 Apr '12, 08:38

mauro-1's gravatar image


edited 06 Apr '12, 10:23

I don't know what you mean about "transparency and fairness". The problematic issues raised along the course were the "grading bugs" in the "grading robots", which IMHO affected everybody. I don't see how the robots could be "unfair" or "opaque" - unless there were some kind of "black-lists" implemented, under a paranoid "conspiration theory" scenario, which IMO never happened here.

So, at least specify the non-technical aspects you think can be improved. Any given so far...

(06 Apr '12, 23:53)


Jose's gravatar image

@jasa, you can search for my other posts and comments to comments to get an idea. Like in here:

How many students completed the course. And any chance of one last office hours?

(07 Apr '12, 11:10)


mauro-1's gravatar image

@mauro, your post is just non-sense to me.... please specify the non-technical aspects you are talking about

(07 Apr '12, 11:52)

Alvaro G. Oj...

Alvaro%20G.%20Ojeda%20Martinez's gravatar image

Follow the links. Think. Do your own research. Think. Repeat.

(07 Apr '12, 12:01)


mauro-1's gravatar image

thanks for that, I have wasted 5 minutes of my life reading your non-sense comments. But I confirmed something.....You should go check yourself my friend... you might have some mental problems....

(07 Apr '12, 12:12)

Alvaro G. Oj...

Alvaro%20G.%20Ojeda%20Martinez's gravatar image

So fast? I'm afraid that's not enough time for thinking and doing your own research, Alvaro.

If you read their (recently added) Privacy Policy, they mention that they are enabled to sell our Personally Identifiable Information along with the company in the future, without consent. In other words, information about you and me is now part of KnowLabs Inc's value. They are building value based on that, that is, they are turning you and me(or information about you and me, which is essentially the same), and about every other student, into a commodity.

I don't know about you, but I don't like being treated like a piece of linen. I will not take any more courses with them, and I suggest you to do the same. I'll also politely ask them to remove all the Personally Identifiable Information they have about me.

(07 Apr '12, 13:24)


mauro-1's gravatar image

I hear you, Mauro.

On the internet, generally the rule applies that if you're not a paying customer, you're the product.

And it also doesn't necessarily make business sense to disclose how you make a profit from something you offer for free on the net.

The folks running udacity are making things up as they go, because progress is too fast to plan everything in advance. An important test will be how transparent they are about sharing potentially valuable information, and how much they keep as trade secrets. I'm watching but not judging yet. What I've seen so far is excellent, and better than existing alternatives.

(07 Apr '12, 15:12)

Christoph M

Christoph%20M's gravatar image

I completely agree with you @Christoph...

(07 Apr '12, 15:55)

Alvaro G. Oj...

Alvaro%20G.%20Ojeda%20Martinez's gravatar image

I don't like being the product, at all. I'm a person, not a commodity.

They present themselves as revolutionizing education, and they are just merchants. That's what I dislike the most. That they are deceitful.

And I think all this is very very damaging in the end. Much more damaging than pepper spray. I'm refraining now. Good bye.

(07 Apr '12, 17:54)


mauro-1's gravatar image
Your answer
Question text:

Markdown Basics

  • *italic* or _italic_
  • **bold** or __bold__
  • link:[text](http://url.com/ "Title")
  • image?![alt text](/path/img.jpg "Title")
  • numbered list: 1. Foo 2. Bar
  • to add a line break simply add two spaces to where you would like the new line to be.
  • basic HTML tags are also supported



Asked: 25 Mar '12, 21:57

Seen: 1,631 times

Last updated: 07 Apr '12, 17:54