These are chat archives for FreeCodeCamp/DataScience

15th
Oct 2016
Zarak
@zarak
Oct 15 2016 07:48
I was just reading through some earlier messages. The website for Coding the Matrix has all the stencil files as well as a submission script for an autograder, but I'm not sure if it still works.
There is also this course on edX for teaching probability in a more computational way.
Zarak
@zarak
Oct 15 2016 07:54
By computational I mean implementing your solutions in code :)
evaristoc
@evaristoc
Oct 15 2016 10:58

@cuent Sorry, my fault: I was mentioning d3.js because in the presentation Evelina Gabasova used F# code apparently calling the t-SNE resulting plot by using a d3.js-like library? No: d3.js can not be used directly to prepare a presentation: it could be as embedded code for charts or similar. Sorry for that.
@cuent : again, nice the youtube link you gave us: we have been discussing some topics mentioned by her here. I think we can continue in that direction so we can have this channel as an additional reference also for that kind of discussions! Also interesting other GoTo Conference talks.

@erictleung Thanks! Have you experienced its use with some data? Share with us and specially with @cuent who is the person who first asked about!

@zarak here you go! excellent, thanks for the info!

@jacqueline-homan, @jomcode, @razerh0 @darwinc : you brought the F# to this chat and now you managed to persuade some of us to start something with F#. Any place to start? We can list from your previous posts:

Can you guide us? Please share!

Question: I am finding that my 4-core CPU and 16G Ram are limiting my possibilities. Any suggestion for someone who would like to use cloud? Any suggestion about GPU in cloud too?

CamperBot
@camperbot
Oct 15 2016 10:58
:cookie: 310 | @zarak |http://www.freecodecamp.com/zarak
:warning: could not find receiver for darwinc
:cookie: 116 | @jacqueline-homan |http://www.freecodecamp.com/jacqueline-homan
:cookie: 327 | @cuent |http://www.freecodecamp.com/cuent
evaristoc sends brownie points to @cuent and @erictleung and @zarak and @jacqueline-homan and @jomcode and @razerh0 and @darwinc :sparkles: :thumbsup: :sparkles:
:cookie: 477 | @jomcode |http://www.freecodecamp.com/jomcode
:cookie: 49 | @razerh0 |http://www.freecodecamp.com/razerh0
:cookie: 419 | @erictleung |http://www.freecodecamp.com/erictleung
karansaulakh
@karansaulakh
Oct 15 2016 11:16
hello campers! Where can I learn data science from for free and easily?
evaristoc
@evaristoc
Oct 15 2016 11:55

People

Designing Information - Information Types: there is more than one...

I am currently reading a book about Information Design and wanted to share some of the concepts with you. The book start by citing a historical reference to a quote attributed to Julius Caesar suggesting him as the person who said there were three types of information. His views were certainly politically driven but the classification deserves consideration by anyone who want to convey information to the public. According to the author of the book (J. Katz, Designing Information), there could be more than 3 classes to include other ways of "Information":

  • Information = Facts disclosed to the audience
  • Uninformation = Decoration of the Facts disclosed to the audience; according to the author it is here where the designer plays (OBS: not sure about this claim as the author suggested, but certainly playing with the "Decoration" of the Facts should actually amplify the effect of either information or non-information)
  • Noninformation = Decoration that is made appears like Information when it is not, but that not necessarily mislead but rather confuse or distract the audience from the actual Information
  • Misinformation = Non-facts, Lies masqueraded as Information; not necessarily deliberate, but always misleading; eg. wrongly analysing incomplete data as complete
  • Disinformation = deliberate Misinformation

I should add this concept for or change names to include:

  • Undisclosed Information = Facts that are not disclosed - it is a way of Uninformation to me, but not necessarily misleading when not of the interest of the public or because must be kept confidential

I would also add IMO:

  • Hidden and Unhidden Information = talks more about the intention of the Information and how deep you should go to reveal the Fact that it is intended to be reported; more to do with Communication probably (other topic...)

@karansaulakh have you seen this article?
https://medium.freecodecamp.com/if-you-want-to-learn-data-science-start-with-one-of-these-programming-classes-fb694ffe780c#.zfqj3rkpe

You can also contact the author to ask for more. We also have a list of suggested sources through this channel but we need to collect and organise them. Possibly a task to be done before the end of this year.

karansaulakh
@karansaulakh
Oct 15 2016 12:18
thanks @evaristoc hope it helps
CamperBot
@camperbot
Oct 15 2016 12:18
karansaulakh sends brownie points to @evaristoc :sparkles: :thumbsup: :sparkles:
:cookie: 317 | @evaristoc |http://www.freecodecamp.com/evaristoc
evaristoc
@evaristoc
Oct 15 2016 21:49

People

Draft of a viz about topics in the FCC Main Chatroom

As you might know I have been working on ways to extract topics and summarise information from the chats. Recently I have exploring some already known techniques. After comparing their outcomes I decided to use Non-Negative Matrix Factorization for being:

  • easy
  • good enough for my computer capabilities
  • effective

The results, albeit the usual subjectivity when analysing human behaviour and specially language, were actually acceptable.

Also you might have noticed that my interest would be to combine message and social networks. So I did an analysis also based on who talk about the possibly revealed topics.

I made a draft exploration of only one week and 6 extracted topics out of 20 possible ones. Usernames are given but the week analysed is not directly quoted. You can find a first visualization of this project in my codepen
http://codepen.io/ecccs/pen/yaxyEa

I will be open to your feedback.

Amelia
@apottr
Oct 15 2016 21:52
@evaristoc what are the seemingly random letters littered about for?
evaristoc
@evaristoc
Oct 15 2016 21:53
@apottr sorry which random letters?
Amelia
@apottr
Oct 15 2016 21:53
Screen Shot 2016-10-15 at 5.53.11 PM.png
Letters scattered about
also it doesn't seem that the usernames in the bubbles change, just the colors of the bubbles
evaristoc
@evaristoc
Oct 15 2016 21:54
Ah! those are truncated usernames. I decided to leave them because I was a bit lazy to delete them and also because I personally found it nice: gave me the sense of amount.
Amelia
@apottr
Oct 15 2016 21:54
I see
That's what I assumed I just wasn't sure whether it was the case
evaristoc
@evaristoc
Oct 15 2016 21:55
But I agree it doesn't look clean, yeah... I personally like it so.
Amelia
@apottr
Oct 15 2016 21:55
That's fair
When was this data collected?
evaristoc
@evaristoc
Oct 15 2016 21:56
About... 1 month ago?
CamperBot
@camperbot
Oct 15 2016 21:56
you need to ask about @someone!
evaristoc
@evaristoc
Oct 15 2016 21:56
Maybe more...
Amelia
@apottr
Oct 15 2016 21:57
Ahh, that explains why I don't recognize any of these usernames haha
evaristoc
@evaristoc
Oct 15 2016 21:57
No more than 2 months ago...
:) :) Sure! The data I am showing is from last year. I collected more data that will be part of a more complete project. Showing information for about one year data.
Amelia
@apottr
Oct 15 2016 21:59
It seems like the only thing that really changes is the colors, is that intended?
evaristoc
@evaristoc
Oct 15 2016 22:00
@apottr but hey! SaintPeter is a legend... and qmikew1 I think is still active in the chat? I haven't heard from jbmartinez.
@apottr for now yes. I was planning to change sizes too but I then realised I didn't prepare the file to get that info. It is still a work-in-progress. But what else would you do, if it would be you?
Amelia
@apottr
Oct 15 2016 22:02
I would try to improve load times between topics but other than that I think it looks good :D
evaristoc
@evaristoc
Oct 15 2016 22:07

:+1: yes thanks! Also I must admit I am not a rock star in css+html, but I am working my way into it. Let's see how it goes.

@apottr If you come with ideas, observations, or things related to your interest, please share? We certainly want to know. Thanks for the feedback. Going to bed. Take care!

CamperBot
@camperbot
Oct 15 2016 22:07
evaristoc sends brownie points to @apottr :sparkles: :thumbsup: :sparkles:
:cookie: 402 | @apottr |http://www.freecodecamp.com/apottr
Amelia
@apottr
Oct 15 2016 22:07
Gnight!