Show HN: Moocha.io – A search engine for MOOCs | Hacker News Reader

Show HN: Moocha.io – A search engine for MOOCs

moocha.io

231 pointsmxeven54 comments

You might want to check out https://www.class-central.com which offers search as well as reviews.

I always begin my search with Class Central. It's so useful!

I gave Class Central a try and searched for Kubernetes. First, it took quite a while to load, up to the point I was thinking the site was down. Then I had difficulty finding the tiny search button located at the top. When I started typing, a list of courses started appearing, but hitting the blue button List All returned me back to the home page. I tried for a second time and hit Enter instead, which presented me a list. Then a very annoying Popup appeared that I could not dismiss for at least 10seconds how much I tried hitting the X. This popup later appeared several more times, then the dismiss worked immediate. The top listed item was an advertisement that was a. not related b. almost not discernible as an advertisement. I clicked on the third item and came to this page https://www.class-central.com/institution/linuxfoundation which wasn't what I looked for, I probably hit the wrong link with my finger as links are very close to each other. Went back and now more carefully hit the correct link. It looked interesting so I hit the big green button only to be greeted by my browser that This page was blocked because of a filter, at which point I gave up.

I suggest Class Central that they hire themselves a proper UX designer.

"Moocha" is the colloquial Tamil word for urine. I giggled like a child.

Input/Output

Seeing as we are off topic...

It made me think of cow's milk, coffee, and chocolate. I prefer your mental image and it also made me chuckle.

Sort of on-topic, trends in naming things confuses me. Are we using so many strange, to me at least, names for things because there are so many products that we are running out of unique names to use?

I also don't see many descriptive names anymore. Joe's Automotives tells me that it's probably a business about cars. Even Microsoft tells me that it is software for a microcomputer. Moocha at least has MOOC in it, though I had to look up the definition of MOOC.

Birds of same feathers. The same thought and same reaction moocha.io. Next up aaya.ai.

Wow the speed is truly awesome - 125ms for the main page and 169ms for the search.

I wonder what's the backend of this app. BuiltWith doesn't detect much.

Edit - I really like the simplicity of search box (no JS):

    <form class="search" action="/search" method="get">
         <input name="query" placeholder="Search" type="search" required>
         <button type="submit"><svg> ..search icon.. </svg></button>
    </form>

Seriously. It was so fast I was convinced it was precaching results as you type but it's not. Would love to read a write-up on this.

I thought you and the other comments were exaggerating, but you're actually not. This feels impossibly fast from a user perspective. I'd love to see a write-up of this as well.

Flask for web framework, Elasticsearch for search. The speed might be due to the fact that there is no JS / images on the front end.

Just had a conversation last night about SQLAlchemy support of NoSQL dialects like. I bet you didn't use a Model layer. Right? Is there anything you want to share in that regards?

There's a distinct lack of Javascript, few (no?) images, and the CSS seems very light. That'll get you a long way right there.

[EDIT] for a similar "oh, I forgot the web could be this fast when we don't pile a ton of shit on top of it" experience, try Basic HTML gmail some time. Its full-page loads are (most of the time) much faster than AJAXy Gmail or Inbox. See also: craigslist.

How does https://whosmentoring.com compare? Albeit hosted in SG so may have more latency.

Never mind. The site loads now. I was a bit confused about the site. As in what kind of mentoring are the listed people providing. The HN link at the bottom helped getting a better picture: https://news.ycombinator.com/item?id=15148619

Sorry I haven't enabled s yet.

Should be

http://whosmentoring.com

Had mixed results on my first two searches:

"C#" returns C++ and Objective C classes, several ranked above actual C# classes.

"Unity" as in the game engine, returns classes on "unit testing".

How are you approaching stemming?

Similar problem here: searched for "CFD" expecting to see hits related to Lorena Barba's excellent intro-level fluids courses and instead got hits for "CAD", "CFT", and "CFP".

Helpful work. Simplistic UI and fast results. Would love to see some more websites like khanacademy, udemy etc in the results too. Some filters like cost/starting in/difficulty level etc. will be helpful in making an informed decision and I am sure data should be available from the original course page. Best of luck.

You might want to consider adding filters. We built an online MOOC directory, specifically for R courses (http://www.r-exercises.com/r-courses/), and noticed that the majority of visitors use the filters we offer. E.g. finding all courses that are free, aimed at the beginner level, or offered on a specific platform.

Interesting. I will definitely consider adding filters in the future.

Originally, I just wanted to a build minimalist search engine that just gave the bare essentials (kind of like Google?)

Another filter I think is very important: only courses that have video lectures.

I like the minimalism. Filtering is great, but more valuable when you know exactly what you want. That said, a (very) few filters for the most common reasons a searcher knows that they require would probably be helpful. Maybe start with just a free/paid filter?

Another filter: language. Many people don't speak English. Maybe you can automatically detect the language and add to your list of languages as you go?

Not sure the search algorithms are working as expected. Consider: https://ibb.co/iGQQPQ

or https://www.moocha.io/search?query=basket+weaving

or https://www.moocha.io/search?query=ethereum

Haha, yeah. I'm still working out my Elasticsearch queries. I thought it was working well enough to at least get some feedback though.

You also can try https://coorsy.com

The search results in this one are much nicer, including the image, course start date, cost, and current enrollment count. OP, would be good to take a peek.

https://coorsy.com/search?query=r%20data

https://www.moocha.io/search?query=r+data

The course information in the coorsy results is valuable, but stock photos are a waste of space on a search results page.

Not a bad start at all. I've completed a lot of MOOCs in my time and have been on the content creation side for online learning as well and it's great to see another tool help with discovery. I think you should add some featured/random courses on the homepage, suggested categories would be good too. I searched "tech" to see what was on offer and got some less than desired results with many of them being about teaching with only some results there for technology based topics.

Overall good start and a useful tool. It will be good to see it developed further.

Being a sucker for MOOCs, this is pretty exciting.

The search algorithm doesn't seem to work right, though. Also, a bit more information would be nice, i.e. when the next offering starts, if it's paid or free, etc.

Great idea and hopefully it will help connect a lot of students to some great material! I noticed it wasn't indexing https://www.kadenze.com. Is there any chance you could add these to the listings? We are focused on arts education primarily but also offer highly technical courses, such as ones in Deep Learning, but with a practice-based mentality towards teaching. Let me know if I can help at all!

Nicely done, might use that from time to time!

I'd reconsider the violet background, it's a bit too agressive and I almost closed the page immediately.

Could someone explain how this works? Meaning, I want to understand how the application uses the data from several MOOCs and get the result. Is that they manually build the data store and indexing on top of it? If so, what happens if a new course is added to the site? Link to source code would be really great, if you can provide one.

While we have a lot of stuff out of date and FB login is not working at the moment...but...

will plug https://www.coursebuffet.com

The aim is to group courses from different providers as if they were all part of one university and you can follow a curriculum.

Feedback for the authors: Make it about planning. There are so many great courses, and no time to go through with all of them. Bookmarking & organization features can capitalize on a huge oversight by the MOOC providers (i.e.: there are no bookmarking features on Udacity & Coursera).

Great job! I hope'Popcorn time for MOOCs"is coming out soon too.

Nice, please also add NPTEL: http://www.nptelvideos.in/

Many useful technical courses on there of high quality.

How is there not any Khan Academy results in any of the searches?

I haven't added Khan Academy yet. However, that should change soon...

The autocorrect/fuzzy search is a little too aggressive.

For instance searching for "PHP" or "REPL" yielded results with "PHD" and "Real..."

How are the results ranked?

It would be nice to have possibility to specify the course starting date like 'in 7 days', 'in 30 days', etc.

How do you index the courses? Use an API per provider? Or scrape the pages and do machine learning?

Thank you so much for this wonderful piece of work.

Please keep improving it.

How many users does your search engine Moocha.io have ?

You might want to check out https://elearning.club. It offers filters by cost, duration and type (eg. Course, Conference presentation and Shows)

The point of a Show HN is for the author to get feedback on their project. It is not an opportunity to sell/recommend a competing product/tool. I am starting to see this anti-pattern a lot in Show HNs, please stop!

Is it? I just glanced the guidelines and FAQ (only on my phone) and didn't notice anything specific to this. Genuinely curious if that structure exists.

Also pointing out competitors might be some of the best advice one could give IMO.

I think there are many ways you could point out a competitor. If a competitor did something different that's notable I think it makes sense to share and explain how the OP could learn from that.

Sometimes 1-liners like "Checkout xyz, been around since 2014" seem a little like an ad leaching off of the original Show HN.

I think it depends strongly on the content and the framing of comment.

Agreed - part of getting feedback on a project is being shown other, similar projects that you can learn from.

I agree that Show HN shouldn't be treated as an opportunity to market another product/project as an alternative. At the same time, being pointed to similar projects that interested users are telling you do something well is highly valuable feedback - in this example, it seems that a common theme of recommendations is the presentation of results. That seems like valuable for the creator.

Reply on news.ycombinator.com