These are chat archives for FreeCodeCamp/DataScience

Aug 2016
Eric Leung
Aug 22 2016 04:22
If you're familiar with the Gartner hype curve (basically a curve describing typical stages of integrating new technologies into the world), they've apparently recently added machine learning to it.
Aug 22 2016 08:35
@erictleung :+1: !
Alice Jiang
Aug 22 2016 12:50
4D printing? How do you print time? Cause I will buy up every bit of it that I can get my grubby hands on
Albert Jonathan
Aug 22 2016 15:58
@evaristoc +1 I see the same trend in high-end web hosting and cloud platform.
Aug 22 2016 20:25
yes in fact it seems an overall trend... and as long as it is "standardisable", it is easy to let it to a single host to deal with it... it is the strong incusion of well supported huge possible standards for data analytics / big data / data science what I find fascinating in this case... I wonder about the final outcome and the overall impacts... in fact the emphasis made by those big companies will IMO certainly affect even more the perception that we already have about data, analysis, etc... and how to access it. IBM and MS (or eg. Oracle with MySQL) storming into the Data Science Open Source territory in the way they are doing now will bring a lot of interesting surprises...

Aug 22 2016 20:36


For those taking the Spark Course, Data Analysis

I just finished the videos of this week. It is my guess that the lecturer is suggesting to work with Singular Value Decomposition (SVD) technique. I have used that for comparing and deciding similarities between texts for previous projects here (for text analytics SVD is the most known Latent Semantic Analysis method). SVD is a first elemental choice method for the application the lecturer is suggesting too (recommenders).
In the last video of this week of the Spark course, the lecturer didn't appear to mention we could be working with SVD: he instead briefly describes the ALGORITHM to make it works. SVD is in any case a matrix transformation, and that is one of the reasons why Spark is much better than Hadoop-based technologies: algorithmically you require some recursive-like behaviour for those transformations which is cumbersome, if not impossible, to implement in Hadoop.
See you around!!
Aug 22 2016 22:55


Statistics for Sport Statistics (NBA): Difficult for no statisticians...

Excellent work the following, and very well explained, but you really must know statistics to understand the details, hard to follow...

Like his remark when the presenter affirms they started doing big data but the actual valuable job was done on small data...

Let me know if you have questions about this one...