thegurus

That diagram in the second picture is my Python setup

Hi guys! Released this post last week in my blog. It is a curated list of excellent Data Science related courses I took over the last 5 years and found them worthy. If you are into DS or transitioning, it may also be helpful for you. Thanks for reading!

David.

Hi guys, wrote this post about how to get a decent Data Science environment. Appreciate your feedback :-)

Hi guys, this is the repository of a workshop we tought last weekend in Madrid, maybe you can find it useful. The content is:

* How to setup MLFLow, a tool for ML experiment tracking and model deploying, from zero to hero. * How to track ML experiments with MLFLow * How to put models to production with MLFLow. * How to deploy models to production in AWS Sagemaker with just a couple lines of code. * How to setup Apache Airflow, a powerful tool to design, schedule and monitor workflows. * How to create workflows that take advantage of deployed models.

Hi HN,

As promised, we open sourced the app powering https://datatau.net, a Data Science newsboard based on HN. To give it some context, original site (in .com domain) has been down from May 15, so we cloned it and continued the project (got no answer from original maintainers of the site). We published it on HN last weekend and made to the front page for some hours.

We have no previous experience with open source projects more than opening some pull requests, so any help from you guys is very welcomed.

In case you want to start a community based on HN style and feel (and do not like alternatives like lobste.rs for some reason), this repo can be a good starting point. Not all HN functionalities are implemented and there are lots of things to polish and improve, but the core is there.

We did it. Got no answer in a month.

another try to comment edited

Wow, it derived in a meta discussion about HN/Lobste.rs.

We have been talking about implementing it, but our final decision is sticking to the original code and make it open source asap to make it better.

Our main reason is that lobste.rs is written in Rails and we would have to learn a new framework to adapt it to our needs, while Python (Django) is more extended among the Data Science practitioners (main target of DataTau and potential contributors).

An open sourced Python alternative may also be interesting for many people.

Yes, it is very unsemantic. The front is made of tables of tables of tables with lots of deprecated tags and most of the styling is made in the html file instead of a stylesheet :-D. We copied the original HN/DataTau frontend with little changes. In the future, we would like to improve it, keeping the HN simple / old fashioned front

You are right. Django default auth backend handles this perfectly, but we wanted that functionality of login + signup in the same form, and that signup autofill in case login fail, so made our custom auth views and screw up some things (like username and password validation and going back to previous page after login)

It was an already existing product and we wanted to stick to it as much as possible!

After we resolve all bugs we want to opensource it. If you give us your email we'll let you know when it is available. Here is our GitHub page: https://github.com/TheGurus

It is Data Science! You are not the only one asking :-D

Sorry guys, do you like ^[A-Za-z0-9@#$%^&+=]+$ for your passwords?

hahaha you're right! we've just setup the ^[a-zA-Z0-9_.-]+$ regex for both username and password. Thanks for the pen-testing

Wow, to be honest we had no idea that it even existed. Thank you for the info, I imagine the idea was cloning: https://github.com/lobsters/lobsters and setting up with DataTau domain and custom .css to make it feel like HN

Do you mean the frontend? In this case we choose to make a copy of HN because we really LOVE its retro frontend :)

Thank you for discovering this bug! We are (hot)fixing it. Had to delete your user though, please sign up again (with a normal username if possible :-D).

Yes, the main goal of this new edition is a 24/7 up site with the same content :) We have in mind posting some old content since it can be recovered.

We are working on it. Wanted to launch asap to give the community a sense of continuity since the original DataTau has been down for too much time (since May 15), so had to give up some features for this early version.

We have to: - fix bugs - improve backend, algorithm to assign karma and sorting posts in frontpage - implement a newsletter with best posts - make an API - opensource the code when it is a bit more polished