I want to be a

harvard_lib

These are my main issues with higher education today:
-it divides society into winners and losers along socioeconomic lines
-it delivers education through a single outdated model
-it’s an inefficient way of getting people into the jobs we need them to do right now

For all of these reasons, we need to update our thinking on how people obtain the skills needed to contribute to society in their chosen occupations. ‘Going to college’ has been a decent solution for this in the past, but we can do better now.

We can separate education from the cost of moving away, living in dorms, and going to lectures – allowing people outside the privileged middle and upper class to obtain recognized higher education. We can move away from the watch-lecture-take-notes system of delivering education to one that allows students to learn at their own pace with their choice of media. And we can directly connect individual employable skills to requisite training without bundling that training with the overhead of a college degree.

So how do we start to move from where we are now to a more efficient and equitable system of higher education?

1) We help job-seekers find out what skills are needed for the types of occupations they’re interested in. The skills that employers are looking for can be quantified; you don’t need to guess what kinds of things will be useful to learn.
2) We show where those skills can be obtained online for free. There are lots of resources out there; most people don’t know about them.
3) We come up with better ways for employees to prove to employers that they have the skills that they say they have.

Thanks to some nice tech solutions, we have more efficient, more distributed, and more open ways of getting a cab, renting accommodations, and paying for things. I think we can similarly improve education and the labor market with the right technology. To demonstrate, I put together a little web app that addresses the first two points mentioned above.

-The app gets the user to enter a job title/profession that they’re interested in.
-Then it searches for those jobs on a popular jobs site.
-It looks through all the job postings that it found and grabs the most commonly requested skills.
-It presents those skills with links to free online educational resources covering the topic.

Go ahead and try it for yourself (best viewed on a desktop) with ‘database developer’ (it could take 10s for the results to load). You’ll see it picks out things like ‘SQL’ (the most popular programming language for interfacing with databases), and Java (another programming language), and some other database related stuff. There’s definitely some very useful information there. But it also picks out the word ‘using’, which isn’t good. You’ll get a lot of almost humorous results for most jobs, but tech related stuff (try ‘java developer’) seems to be alright, probably because it’s easier to define required skills in that industry.

There are definitely things that could be done to improve the quality of the results, but that’s not the aim here. I’m really just trying to demonstrate the conceptual link between jobs, skills, and education that I feel is lost with our current system. Hopefully you see the potential.

“I want to be a _______. These are the skills employers are looking for right now. This is how I can learn those skills right now for free.” That’s what we should be striving for. The system that we currently allow to exist is so far from that ideal, and it really doesn’t have to be that way. There’s a lot of cool things going on in ed tech right now, and I think the more people are aware of how beneficial they are, the better chance we have of bringing about the shift in thinking needed for them to be socially accepted.

PS: I think point 3) above is the next big challenge. The resources for people to learn outside of a university setting are already pretty well in place (though we might still need to do some work on helping people figure out the best things to be learning). But how does someone demonstrate to employers that they have learned something? I think the answer is project-based proof-of-knowledge, and open-sourced tests that can be done online without live human supervision. More to come on that someday.

APPENDIX: Software

It’s built using the Python web framework Flask.
-I searched a bunch of random jobs on a jobs site, parsed the text, and compiled a list of words and the frequencies in which they occur in that set of job posts. This gives us an idea of how often these words typically occur in a job post.
-Then I look at job posts for the job title that is being searched and do the same thing. So now I have two dictionaries of words and their frequencies. The ‘specific to the search frequencies’ and the ‘typical job post frequencies’.
-Then I look at the difference between a word’s ‘specific to the search frequency’ and the word’s ‘typical job post frequency’. If the ‘specific to the search frequency’ is higher, it’s probably an important word specific to that search. I look at the words with the largest differences between their two frequencies and consider them ‘skills’ and add them to the list.

There’s a little more to it than that, but those are the basics. The code is up here if you want to take a look yourself or fork it. There’s lots that could be improved: looking for ‘skills’ that are more than one word long, filtering results by things you think are actual skills (maybe from a database of skills from LinkedIn or something), caching results or multithreading to reduce loading time, and matching up the results with more relevant free online educational materials.

The app searches through jobs on Indeed. Shout out to them for having an easy-to-use public API.

Leave a Reply

Your email address will not be published.