mrborgen

Sure, I asked at Reddit/r/machinelearning:

https://www.reddit.com/r/MachineLearning/comments/4kcrqp/loo...

A couple of people contacted me, and the first one I gave a shot was actually a perfect fit. Maybe I was a bit lucky right there.

Hey, I'm the author of this article. Happy to answer questions if anybody have any.

Yes, Datasets.co. A site for sharing/indexing machine learning datasets.

I actually asked this a few months ago. Here is the thread: https://news.ycombinator.com/item?id=11229700

Your assumption is right, in terms of business cost, we'd rather see a false positive than a false negative.

I don't think the scikit learn algorithm differentiates between the two types of errors, in terms of cost.

Though it seems to give less false negatives than false positives overall, when testing on new datasets.

I'll put in the f1 score in the article when I have time.

2000 samples is what we needed to get going yes. We're also noticing variable accuracy on different real world dataset. Meaning the algorithm is biased towards our current dataset.

Getting more data is one of our top priorities going forward with this project.

Author of article here. Happy to answer questions if anybody have any.

Not available in my App Store (Norwegian). Why not make it international?

Sorry for the self promotion, but I wrote a few React tutorials with someone like you in mind. They're only covering the basics of React, coding directly in an html file, so they don't involve getting up and running with Node+React. But I think that should be tackled after you know the basics of React.

Building Your First React App: https://medium.com/learning-new-stuff/building-your-first-re...

Building Your Second React App: https://medium.com/learning-new-stuff/building-your-second-r...

React.JS in 8 minutes: https://medium.com/learning-new-stuff/learn-react-js-in-7-mi...

Other than that, I always recommend the official tutorial: https://facebook.github.io/react/docs/tutorial.html

... and Egghead.io's videos: https://egghead.io/series/react-fundamentals

I like this idea. If properly executed, it'll give people a whole new type of liquidity. I see the App Academy's model (free bootcamp, but takes recruitment fee from companies who hire their graduates) as a very early prototype for this kind of model.

Awesome list, thanks!

It's avtually 19, I just didn't understand all of them to be able to write proper descriptions.

Hey, on datasets added from now on, this'll be an option. Check it out in action here (displayed under the image).

http://www.datasets.co/dataset/All-UFC-Fights

Was that what you were thinking?

Was a bit quick comment last time. Yes, having the amount of observations is a very natural feature to add. Shiuld be fairly simple as well, so it'll be added soon!

Hey all!I built this site because I think it should be easier and more fun to discover new datasets. I'd love hear your thoughts, suggestions and critisism!

Thanks! Would you also be interested in having a 'data preview' feature, where you'd see a tiny part of the data in the browser?

Thanks for the input! The variables are already described, if it was the features you were thinking about?

Great! The search on Apple's own app is just horrible. I'm gonna use this one for sure. Well done!

We are very much considering opening up the chat logs!

Thanks, we're hoping to do something in the same lines. However, it's more tricky when it's a chat, as the answer isn't as easy to present to the whole community and gain more karma from.

Any thoughts on that?

Thanks!

Hey, awesome service! I'd be interested in using this at my site (BugRex.com). We're a chat based help line for developers in the need for help. Sharing code is an issue we haven't solved properly yet. Let me know if you want to chat.

Not very strict demands. If it's still growing and does exciting stuff, I'm interested in hearing about it. (Though that probably is an insane amount of companies)

Interesting idea regarding scheduling. That would require a bit strickter rules for our experts though. As for now, one of the benefitd the experts like is that it's 100 percent flexible. You answer chats only when your'e interested.

Would love to hear about those human decision theories, got links?

That remains to be seen! Depends on the growth, definitely.

Thanks for your input! The are a ton of challenges ahead, and scaling experts is one of them. But as for now, getting demand is harder than supply. Although the demand is surely affected by the quality of the experts.

Great input!

However, it's quite hard to get that kind of product off the ground. Easier to start with a niche and work our way out to broader subjects.

We actually started with ReactHelp.com, then went wider to HTML,CSS and JS. And now, we have Ruby, Python & Java experts as well, so it's going in the direction of more subjects.

Hey, author of the article here. Let me know if you have any comments.

Thanks, it's now fixed!

Hey, we have four.