These are chat archives for FreeCodeCamp/DataScience

30th
May 2016
evaristoc
@evaristoc
May 30 2016 10:52 UTC
Hi everyone:
This is my last progress with my project... if anyone can help me with the css it will be very appreciated:
http://codepen.io/ecccs/full/YWzJOP/

The project doesn't look in codepen as in my computer (firefox browser)! In my computer looks far much better. In an old version of Chrome it also looks fine but the comments are not placed to the right, as in my firefox. It is meant to be responsive and interactive...

Again:

Any help with css will be welcome

(no my stronger point at the moment...). You might need to modify the project locally instead, as codepen might not be the host of the projects...

Emil Aasa
@emilaasa
May 30 2016 13:05 UTC
Ya'll found jesus yet?
I mean, have you started using Docker yet?
Ideal for replicating more complicated development environments across multiple devs!
evaristoc
@evaristoc
May 30 2016 13:09 UTC
@emilaasa :) :)
Emil Aasa
@emilaasa
May 30 2016 13:14 UTC
Honestly it's cheating! Removed most of the work from our sysadmins visavi the dev department
=)
Koustuv Sinha
@koustuvsinha
May 30 2016 13:23 UTC
@emilaasa vagrant vs docker?
your inputs
Emil Aasa
@emilaasa
May 30 2016 13:39 UTC
@koustuvsinha It depends! :D
But for my use case I like Docker
Koustuv Sinha
@koustuvsinha
May 30 2016 13:40 UTC
can you briefly explain where docker scores over vagrant? :P
Emil Aasa
@emilaasa
May 30 2016 13:42 UTC
Well Vagrant is a more complete solution
Run Docker on Vagrant and you're set :P
Koustuv Sinha
@koustuvsinha
May 30 2016 13:42 UTC
like Vagrant is a full VM and Docker is a lightweight one?
Emil Aasa
@emilaasa
May 30 2016 13:42 UTC
Vagrant is a VM-manager
so you can do anything you want there
It all depends on your teams enviroment
Koustuv Sinha
@koustuvsinha
May 30 2016 13:44 UTC
nice. so previously I was discussing with @evaristoc about the possibility of creating a Vagrant for DataScience, which would have all tools preloaded. Can that be done in Docker as well, or what are the limitations there?
Emil Aasa
@emilaasa
May 30 2016 13:45 UTC
Well, there's not really any overhead to use vagrant + docker
For instance if I were to host a online DS course, I would for sure use Vagrant to make sure every students env was exactly alike, from OS and up
And then Vagrant would take care of everything, including installing the docker tools and perhaps project specific docker files
But docker now works in all OS:s with a little bit of installation
This image is a bit out of date, but the starting time and size pretty much holds true
blob
Koustuv Sinha
@koustuvsinha
May 30 2016 13:49 UTC
Here in FCC we are currently developing various Vagrant packages for easy project development in Python, Node or Ruby. you can check us out in the FCC/vagrant room
Emil Aasa
@emilaasa
May 30 2016 13:49 UTC
Cool!
I think that is a great idea
Koustuv Sinha
@koustuvsinha
May 30 2016 13:50 UTC
So, there is a discussion regarding to size of the VM as well. Since students are from various geographical locations, not everyone has access to good internet.
Thus one priority is to have the initial image size low
Emil Aasa
@emilaasa
May 30 2016 13:51 UTC
You could always push as much as possible down to a docker image, and then have vagrant as an optional wrapper with a separate dependency list if people only want to use docker
there's docker toolbox now for seamless dropping in of docker into OSX and win
Koustuv Sinha
@koustuvsinha
May 30 2016 13:55 UTC
nice! checking it out :)
evaristoc
@evaristoc
May 30 2016 15:21 UTC

@koustuvsinha I think we could start working on the "Radar"-kind project I proposed here. @emilsaasa: I don't know if you are in?

The only thing is: I would start by keeping it simple. I was thinking: either a region (possibly at city-level) or at meetup-name level (for which I was thinking nodeschool to start with).

In the second case we can try to focus also on social network analysis (having meetup members as targets), possibly using twitter information, and perhaps geolocation.

The first case (city-level instead) would require more parsing and text mining to identify the meetups that we are interested (remember I suggested to focus on Javascript-related, but that means: react, backbone, angular, etc and won't be easy...). PLUS I would like to combine always that information with social network identification, making it harder...

The fact is that I don't if I have enough time to start working on the hardest idea right from the beginning...

Koustuv Sinha
@koustuvsinha
May 30 2016 15:23 UTC
awesome idea. We can target any meetup, if Nodeschool then we can track the chapter tweets and associated user interest, repository and sign-up interest
evaristoc
@evaristoc
May 30 2016 15:24 UTC
Positive
Koustuv Sinha
@koustuvsinha
May 30 2016 15:24 UTC
Suppose if you are targeting a bigger event, say JSConf, then you will get loads more data on tweets, interest spikes and geolocations
evaristoc
@evaristoc
May 30 2016 15:24 UTC
That's more or less the idea... Then any learning from the project could be eventually translated to a bigger one...
Good point...
Let's analyse both? I would suggest to keep that one that has less exposure, as it is likely that although might not be of the interest of everyone it will also make the project a bit more "original"
But I might be wrong...
Koustuv Sinha
@koustuvsinha
May 30 2016 15:27 UTC
we can create a framework for analysing city based or regional events / meetups
Nodeschool will be a smaller dataset, but can be effective if all 290+ chapters are taken into account
evaristoc
@evaristoc
May 30 2016 15:27 UTC
I hope we can make a team of 3, excluding @emilaasa (he would take care of a very specific type of data rendering then...)
About the framework yes...
CamperBot
@camperbot
May 30 2016 15:28 UTC
you need to ask about @someone!
Sam Aiken
@SamAI-Software
May 30 2016 15:30 UTC

Templates are pretty much done. Basic styles for bar charts will be H4 and H5.

BarChartsHorizontal_H5H4.jpg

evaristoc
@evaristoc
May 30 2016 15:31 UTC
Ok: let's explore twitter's API a bit more, and also the meetup.com one... We should agree in a model and a stack framework. So far my strongest framework is MEAN (angular 1.x)...
@SamAI-Software: do you want to check our project for a Radar and see if you feel like participating?
Koustuv Sinha
@koustuvsinha
May 30 2016 15:32 UTC
@evaristoc ok. yeah getting the dataset is the first step :P
Sam Aiken
@SamAI-Software
May 30 2016 15:32 UTC
@evaristoc never heard of that. What's that?
evaristoc
@evaristoc
May 30 2016 15:32 UTC
We are going to make a database... as I see it...
@SamAI-Software : check this
Going for quick jogging... and @SamAI-Software: what happened with H3??
Sam Aiken
@SamAI-Software
May 30 2016 15:35 UTC

what happened with H3??

executed :skull:

evaristoc
@evaristoc
May 30 2016 15:36 UTC
oh... :)
Sam Aiken
@SamAI-Software
May 30 2016 15:36 UTC
:cry:
oops, wrong chat
sorry
evaristoc
@evaristoc
May 30 2016 22:24 UTC

People

The DSR Weekly Digest:

Previous Digest

More about Survey 2016, specially @SamAI-Software. Also other aspects that were commented:

  • Support Vector Machines (@erictleung, @darwinrc) and the MIT course link...
  • A mapping library, openlayers (@emilaasa), and the possibility to start a project
  • A discussion about the apparent focus on visualizations in this room, waiting for feedback (@jacobbogers, @bitgrower)
  • Comments from a Data Science meetup I attended, leading to a proposal to work on a kind of "Tech Radar"; @koustuvsinha showed interest in participating, possibly to combine with @emilaasa interest?

(trying a different algorithm to analyse the week...)

Michael
@m-henderson
May 30 2016 22:49 UTC
Can someone help me get started? I want to contribute to FCC on the data science end of things. I don't really know where to start.
bitgrower
@bitgrower
May 30 2016 23:51 UTC

A discussion about the apparent focus on visualizations in this room, waiting for feedback (@jacobbogers, @bitgrower)

Oh ... I didn't know there was more to the discussion ... :) ...

bitgrower
@bitgrower
May 30 2016 23:56 UTC

Evaluating the meetup-related activity could give an idea about the intensity at which the topic is gaining the interest of the people who leave at those cities/countries.

one of my favorite topics ... early trend detection ... :)