A few months ago I was selected to take part to the Bertesmann Data Science Challenge Scholarship Course on Udacity. The scholarship is basically a program in two phases:
- In the first one, 15,000 participants were selected to take part to a 3-months “challenge course” covering the basics of data analytics (basic statistics, introduction to python and to SQL).
- Among those who completed this challenge course within the three months, 1,500 would have been selected and have the opportunity to enrol for free in one between Data Analytics, Business Analytics or Data Foundations nanodegrees on Udacity.
The first phase closed about two weeks ago, and last Saturday we were told who was to advance to the second phase.
In this post I will write down some considerations about the experience so far, the course structure and what I think can be improved.
Full disclosure: by totally unrelated circumstances of life, I happen to be at the moment employed by one company which is part of the Bertelsmann group (I was not when I applied for the program and when I was selected). Needless to say, what I write here reflects my and only my opinions.
- As should be clear by now, I kinda like online courses and data analytics. So it should come as no surprise that I like free online courses on data analytics. Joking aside, the idea behind the challenge is an interesting one, and I wish I had come across something like this a couple of years ago, when I would have probably benefitted more from the materials presented. Despite my limits, I am to a lower or higher extent already familiar with most of the materials that were covered by the challenge course, which ended up being mostly a helpful refresher. This personal consideration aside, I think the idea is a good one and provides an helpful opportunity to many people wishing to have a first bite at stats and programming for data analytics.
- As I mentioned above, the challenge course covered basic topics. It is always difficult to make a selection on what to include and what to leave out from a three-months online course. In general, I have the impression that the amount of work was adequate to the time available, however I also have the impression that digging a bit more in depth with the statistics module, perhaps at the cost of sacrificing the SQL part, would have been a better option. For once, the stats part covered the basics up to the normal distribution, without dealing with statistical inference, which is where the fun begins. Including inference (or introducing simple linear regression) would give the students a better understanding of what statistics can be about, and give some better foundation for working on personal projects (more on this later). The part of SQL, on the other hand, although interesting in itself, is quite irrelevant until one starts to work with (large) databases, which doesn’t happen often to be the case for students practicing on their own. Furthermore, one could gain the same familiarity with the SQL topics with some real – world practice when need be.
- Talking about projects: the challenge didn’t include a final project, although students where encouraged to complete one, possibly in groups. I am not sure if this is the best approach, as the project I took part to was the most interesting part of the whole experience (we tried to see if there is a correlation between salary gap and length of paid of maternal leave at country level, without finding any). More in general the final project tend to be the most relevant part of a MOOC experience, especially those with a vocational-training flavour like Udacity’s. A compulsory final project would compel a student to put in practice what s/he has learned, and perhaps give an additional criteria for selecting students for the second phase.
- The students were accompanied during the three months by some tutors over the forum and slack channel, which were obviously open to all the students. I personally didn’t have any interaction with the tutors, as the materials were pretty clear and instruction straightforward. During the three months I remember only a single instance of a technical mishap, which is something I can live with. I spoke elsewhere about the emphasis that Udacity put on additional resources in terms of coaching and mentorship, and some services were indeed available for the participants, in particular the career portal. However, compared to a regular nanodegree, most of the emphasis was put on the community of students themselves, being the participants encouraged to initiate study groups, group projects and to interact over the forum and slack.
- Talking about slack: I have stated before and I repeat it here, that it is not an adequate tool for a group of 15,000 people. In theory the slack channel is meant to be a “light” and fast way to conduct quick communications as opposed to the forum, meant for more in-depth exchanges. In practice it proves just a cumbersome, confusing and redundant mess. A group of 15,000 people is not a team, and if smaller teams are formed, they have thousands of more practical ways to communicate with. And if we are talking about ease of communication and responsiveness, I think that any inquiry over the forum that required a feedback manage to get at least a couple of answers within 20 minutes. So, no, I still think that slack is not a convenient tool for this kind of endeavours.
- The forum, on the other hand, proved quite a convenient channel of communication and I think that it managed to make the experience of many users much more valuable. As mentioned above, inquires got answers pretty quickly, and as far as I could see, got adequate ones. More than once I was about to post an answer to a question before realizing that already a couple of good answers already got published. There has been a bit of a tendency to share single articles and resources, but then it got more systematized when someone had the idea to set up a database to be directly updated. So, in all I believe that the forum was one of the best tools available for the participants.
- There were also a Facebook group and a website, but I haven’t interacted much with them. The Facebook group was eminently a Facebook group – so mostly to share thoughts, motivational quotes, some resources, and lot of pics of study groups.
So, in a nutshell this is it: if there will be future editions, I would put more emphasis on a final project and on the statistics module, perhaps allowing a couple of weeks more, but other than this I don’t find much to remark.
Udacity stands quite out in the sea of MOOC provider for the emphasis on employability and vocational training, but also in finding alternative ways and solutions in a sector that, although not in its infancy any longer, still has yet to find its place among and alongside traditional higher education providers.
In some ways, Udacity reminds the role of pre-1992 polytechnics in the UK, which despite being less famed than traditional universities, played a much more dynamic and pioneering role in finding new solutions for higher education.
Udacity nanodegrees (as well as similar courses from other providers) still comes with a pricetag that, although affordable in higher income countries, is still out of reach for many people from middle and lower income countries, which would be those to benefit the most from remote providers of good quality higher education. This can perhaps be a further direction to explore for future scholarship paths.