These are chat archives for FreeCodeCamp/DataScience

Jan 2016
James Winegar
Jan 22 2016 03:41
@evaristoc tensorflow is a full blown machine learning framework, not just a library. It has significantly more features than you will ever touch, but in the end they will be easier to implement than a solution based on sklearn and such. Really just a tradeoff of types of versatility at this point.
Nick Robson
Jan 22 2016 07:04
Well, hello, ranking #13 on that scoreboard. hehe (by recent score)
Jan 22 2016 09:56
@nickrobson welcome,#13!
Jan 22 2016 22:52

While dedicating some time to learn d3.js (and the rest: html, css, JS, etc...) I have been wondering when d3.js is superior and when it is an "overkill". For example: I was primarely working on this scatterplot. I must say that when static it was definitively easier in Excel than d3.js. I really started to appreciate some knowledge of the d3.js library after I started adding some interactive features...

The fact is that If you are just working on something static or with pre-loaded capacities, d3.js seems to be too much coding... Even other programs, like python or R, could offer much better and nicer solutions at some point...

Although my main focus in the future will be d3.js, I was still asking myself about preferences between css/html+vanilla JS vs d3.js: when one is better than the other one? I feel I don't have the experience to answer that question. But for those who are more experienced with css/html or even vanilla JS, I found this article very interesting: "You may not need d3" --- the author also goes around a similar proposition for jQuery, suggesting vanilla JS could be sometimes more powerful...