These are chat archives for FreeCodeCamp/DataScience

31st
Mar 2016
Quincy Larson
@QuincyLarson
Mar 31 2016 01:37
@ChristianBertram OK - I'll add Danish
Hey Data Science room - here's my article announcing the survey. If you have a second, please read it and give me feedback: https://medium.freecodecamp.com/we-just-launched-the-biggest-ever-survey-of-people-learning-to-code-cac81dadf1ea
Alice Jiang
@becausealice2
Mar 31 2016 02:24
@QuincyLarson Is the raw data available anywhere to do something like a live update while the survey is open?
Quincy Larson
@QuincyLarson
Mar 31 2016 06:46
@alicejiang1 yes - I can get you the raw raw data
This is for the demographic section (which most people also complete)
Alice Jiang
@becausealice2
Mar 31 2016 06:50
I'm really only interested in the responses country of residence right now if that makes things easier on you :)
OK - good. Gitter didn't attempt to render it.
Alice Jiang
@becausealice2
Mar 31 2016 06:56
Fantastic :smiling_imp:
evaristoc
@evaristoc
Mar 31 2016 15:53

@Mius00 : perhaps you would like to check the following? This is somehow going in the direction of the project I have in mind (just read the summary):

https://www.researchgate.net/publication/220878769_Peer_interaction_effectively_yet_infrequently_enables_programmers_to_discover_new_tools

evaristoc
@evaristoc
Mar 31 2016 17:45

People

About the survey...

Some results are already available at the survey hosting platform (see @QuincyLarson's article at medium to find out...)

BUT we could still do something about cross-referencing the data, which I think is not available at the moment...

For those who are wondering what the hell I am suggesting to do, please check the following link as an example:
https://dc-js.github.io/dc.js/
Roel Verbunt
@roelver
Mar 31 2016 20:39
I created an Academic Honesty tool based on the mongo collection that is backing the API lists. It is a Node script running through all users in the database. It's currently only checking the submitted projects. I'm checking on the following characteristics:
  • -total duration to complete all challenges. Less than 1 day per challenge on average is a hit.
  • -more than 1 completed project within 24 hours.
  • -solution URLs that are not referring to Codepen or Heroku
  • -duplicate URLs as a solution
Any thoughts on what could be checked more?
Victor
@Evaderei
Mar 31 2016 20:41
@roelver Hb submiting someone elses' links as your own?
But then you'd need to store all the links to check?
Roel Verbunt
@roelver
Mar 31 2016 20:42
The result will be a json file like this: [ { "username": "roelver", "onWhitelist": true, "onBlacklist": false, "projectCount": 2, "evidence": [ { "project": "Build a Personal Portfolio Webpage", "completed": null, "solution": "http://r3dm.com", "message": "Project not on Codepen or Heroku." }, { "project": "Build a Voting App", "completed": null, "solution": "http://www.r3dm.com", "message": "Project not on Codepen or Heroku." } ] }]
@Evaderei Thanks for the suggestion. I will think of it.
CamperBot
@camperbot
Mar 31 2016 20:45
roelver sends brownie points to @evaderei :sparkles: :thumbsup: :sparkles:
:star: 428 | @evaderei | http://www.freecodecamp.com/evaderei
evaristoc
@evaristoc
Mar 31 2016 20:46
@roelver there are some people hosting their projects in github.io... I think that is also valid?
Roel Verbunt
@roelver
Mar 31 2016 20:47
@evaristoc OK I'll add github
Brian
@sludge256
Mar 31 2016 20:51
Are there any hits for OpenShift or Nitrous? Maybe even C9 despite it not being an ideal solution.
evaristoc
@evaristoc
Mar 31 2016 20:51

Apart of your list I cannot think of any other additional comparison... I think it will be ok...

@roelver: to get data from the API... is it ready?

@sludge256 C9 perhaps... but Nitrous?
Brian
@sludge256
Mar 31 2016 20:52
Nitrous is similar to Heroku.
Their free plan is pretty restrictive last I saw though.
evaristoc
@evaristoc
Mar 31 2016 20:52
Ahh... ok... no idea...
Roel Verbunt
@roelver
Mar 31 2016 20:52
@Evaderei I have all links stored in the database as part of the user data. So I think I can do your suggested check
Victor
@Evaderei
Mar 31 2016 20:53
@roelver Sweet, any idea how long it'll take each run?
That one check specifically to be clear
Roel Verbunt
@roelver
Mar 31 2016 20:54
Currently I have only 3 users in the database. It is completed in a second :smile:
evaristoc
@evaristoc
Mar 31 2016 20:54
hahaha!
Keep it then like that!
Roel Verbunt
@roelver
Mar 31 2016 20:55
I'm waiting for some more...........
evaristoc
@evaristoc
Mar 31 2016 20:55
Noooo! Don't do that!
The project only compares projects? No algorithms? @roelver?
Roel Verbunt
@roelver
Mar 31 2016 20:58
The tool produces a list of suspects that need to be evaluated manually. I don't want a lot of false positives, so adding a couple of sites to check like openshift, c9, nutrous will reduce the number of suspects.
evaristoc
@evaristoc
Mar 31 2016 21:00
Oh sorry, yes I see...
Roel Verbunt
@roelver
Mar 31 2016 21:00
No, I don't check the algorithms. These are not in my api database, because it is unstructured code.
I could compare the scripts from the FCC user database with others, but that does not necessarily mean that the code is stolen.
evaristoc
@evaristoc
Mar 31 2016 21:03
Indeed, difficult to determine.
Brian
@sludge256
Mar 31 2016 21:04
In regards to the CodePen examples, I was thinking we could add "code canaries".
That way even if people modified their copy and paste jobs to bypass an auditing tool they may not notice the canary.
evaristoc
@evaristoc
Mar 31 2016 21:06
@sludge256 interesting... sort of encryptation or so...
they copy and the resulting paste is encrypted... or inject a virus...
No really: sounds like a good idea for all projects...
Brian
@sludge256
Mar 31 2016 21:08
It's similar to students copying and pasting text for an essay but then changing a few words or sentences so it's harder for instructors to find exact matches on Google or some other site.
Alice Jiang
@becausealice2
Mar 31 2016 21:10
inject a virus.... You have serious issues @evaristoc :laughing:
evaristoc
@evaristoc
Mar 31 2016 21:10
@alicejiang1 in wonderland! yes, I have...
Brian
@sludge256
Mar 31 2016 21:10
Maybe just have it flag the submission ;)
Alice Jiang
@becausealice2
Mar 31 2016 21:11
@sludge256 he'll probably still suggest a virus ;P
Brian
@sludge256
Mar 31 2016 21:12
We can easily parse the code from CodePen also.
https://blog.codepen.io/documentation/api/url-extensions/
evaristoc
@evaristoc
Mar 31 2016 21:12
@alicejiang1 !!! how did you know?????
Alice Jiang
@becausealice2
Mar 31 2016 21:13
@evaristoc I figured out as soon as I heard you try to say "GitHub"
I swear on my life you were saying "you need to jiggle" not "you need to github"
in that moment I said to myself.... This man has some serious issues
Speaking of crazy, I have this CRAZY idea
evaristoc
@evaristoc
Mar 31 2016 21:14
I said hiphop
So you never understood what I said
Alice Jiang
@becausealice2
Mar 31 2016 21:15
That could be said of almost everything you say to me
evaristoc
@evaristoc
Mar 31 2016 21:16
@sludge256 that sounds a good idea... that and a virus for copying the other's work, to give a good lesson
@alicejiang1 hahahaha! come on... that is unfair!
I was under the effect of alcohol...
Alice Jiang
@becausealice2
Mar 31 2016 21:17
Don't drink and advise baby data scientists
Crazy Venezuelan
evaristoc
@evaristoc
Mar 31 2016 21:19
Suorrie (hip)
Alice Jiang
@becausealice2
Mar 31 2016 21:20
What about my hip?
No but I have an idea
You've been workign on using Gitter conversations to help find FCC challenges that confuse people and may need to be redone, right?
evaristoc
@evaristoc
Mar 31 2016 21:22
hmmm.... yeesss...
Alice Jiang
@becausealice2
Mar 31 2016 21:22
Assuming the servers could even handle it, how do you feel about biofeedback?
evaristoc
@evaristoc
Mar 31 2016 21:24
yes... we are a bit far but I could say that for working that in the chatrooms as they are now that could be certainly "easy"...
Alice Jiang
@becausealice2
Mar 31 2016 21:24
Not biofeedback in the chatrooms
evaristoc
@evaristoc
Mar 31 2016 21:25
where then?
Alice Jiang
@becausealice2
Mar 31 2016 21:25
in the challenges. If a camper opts in they have a webcam watching them work on challenges and if they show confusion it flags the problem
evaristoc
@evaristoc
Mar 31 2016 21:26

Oh sorry! I thought it was Natural Language...

I would like to think that it is possible for FCC, but I think it is still to much to wait for that... I think this is the future though...

Alice Jiang
@becausealice2
Mar 31 2016 21:27
Well of course not right now, the servers can barely handle their work load without biofeedback data streams ;P
evaristoc
@evaristoc
Mar 31 2016 21:28
I believe that future platforms like this one will start implementing something similar, specially when Universities solve the dylemma of whether maintaining or not the teaching staff
hehehe!
Alice Jiang
@becausealice2
Mar 31 2016 21:29
You know the saying "those who can't, teach. Those who can't teach, teach gym."?
What will happen to "those who can't" when teaching positions are obsolete?
evaristoc
@evaristoc
Mar 31 2016 21:31
I think there will be a lot of people teaching in the gym? Getting a personal trainer will be cheaper?
Alice Jiang
@becausealice2
Mar 31 2016 21:32
I hope this happens soon, I'd pay for a cheaper personal trainer
evaristoc
@evaristoc
Mar 31 2016 21:35
No to be honest, I have been checking several discussions and fora about the topic... first the arrival of moocs and now... without being too clever I can just see how to use some simple Machine Learning algos to help people coding, for example...
Alice Jiang
@becausealice2
Mar 31 2016 21:35
You think teachers really will become obsolete?
evaristoc
@evaristoc
Mar 31 2016 21:36
The use of AI like emotion recognition systems is now much easier and technologically getting so mature that it wouldn't take long before it becomes pervasive
No in all areas, the role will be re-adapted and yes: you may have less teaching staff eventually for some specific areas...
bitgrower
@bitgrower
Mar 31 2016 21:38

in the challenges. If a camper opts in they have a webcam watching them work on challenges and if they show confusion it flags the problem

...that could be as annoying as clippy ..

evaristoc
@evaristoc
Mar 31 2016 21:38
Or maybe just the assistants...
bitgrower
@bitgrower
Mar 31 2016 21:39
"you want some help? you want some help?"
evaristoc
@evaristoc
Mar 31 2016 21:39
It would more like "I AM SURE THAT YOU NEED MY HELP"
I AM A COMPUTER, YOU KNOW?
bitgrower
@bitgrower
Mar 31 2016 21:40
hahahahahahaha @evaristoc ...
Alice Jiang
@becausealice2
Mar 31 2016 21:40
@bitgrower No one would even see what's happening. It would just keep a tally of how many campers have are struggling with a challenge 100% behind the scenes and the FCC team would decide if a tally is too high and rework the challenge
bitgrower
@bitgrower
Mar 31 2016 21:41
there's a sticker on my webcam for a reason ... I wouldn't want to be webcamming people without very clear &obvious permission from them ... (they could be in their underwear ... or worse...)
Alice Jiang
@becausealice2
Mar 31 2016 21:41
Why would we want something to pop up and offer to help a camper code? that's not how coding works
evaristoc
@evaristoc
Mar 31 2016 21:41
Well: I was already commenting in this same channel some time ago about Learning Analytics and Big Data, and that it is more or less what they are doing...
bitgrower
@bitgrower
Mar 31 2016 21:42
well ... sourcing from webcams vs. text does seem just a tad more intrusive ...
Zac Cassini
@zcassini
Mar 31 2016 21:42
@alicejiang1 many ides do work like that. as do most word processors.
evaristoc
@evaristoc
Mar 31 2016 21:42
absolutely
Alice Jiang
@becausealice2
Mar 31 2016 21:43
okay, first of all if you're going to quote me, read what I said. A user has to opt in. Second don't judge something until you understand how it works. The software doesn't record anything for anyone to use it analyzes what it sees in real time and the data it collects is what is returned
bitgrower
@bitgrower
Mar 31 2016 21:43
I saw Amazon's echo and I went ... I don' think so ...
Alice Jiang
@becausealice2
Mar 31 2016 21:43
@zcassini What IDE are you using that sees you struggling to learn what a for loop is and explains exactly how that works?
bitgrower
@bitgrower
Mar 31 2016 21:44
hey ... I didn't see all your comments, and I suspect you didn't see all of mine ... I am simply stating my concerns about the general concept ... if they don't fit, then fine ... but they weren't intended as a "judgment"
evaristoc
@evaristoc
Mar 31 2016 21:45
@bitgrower intrusive? not really... a bit artificial perhaps... Artificial Intelligence... it is interesting how fashionable such a non-human way of interaction could get... (sorry bitgrower...)
bitgrower
@bitgrower
Mar 31 2016 21:45
despective ???
Zac Cassini
@zcassini
Mar 31 2016 21:45
@alicejiang1 not quite at that level. But many ides offer code completion, linting that pop up and help somone who is struggling with learning to code.
evaristoc
@evaristoc
Mar 31 2016 21:45
But it is the way it goes...
Zac Cassini
@zcassini
Mar 31 2016 21:46
heck some will even help in the writing of for loops.
Alice Jiang
@becausealice2
Mar 31 2016 21:48
@bitgrower You quoted my only explanation of the idea, and I have seen all of yours with regards to this conversation. You are making assumptions about the software and coming to conclusions based on the assumptions, which is what judgement is.
Zac Cassini
@zcassini
Mar 31 2016 21:49
the point being that modern programming tools do pop up and help the user. The idea to take it a step farther isn't at all radical. Not that it couldn't be argued that what we have now is wrong or an unneccessary crutch. Not that I would make such an argument.
bitgrower
@bitgrower
Mar 31 2016 21:49
language is a very imperfect mode of communication, @alicejiang1 -- you seem to have been offended -- I apologize -- however, I differ on your interpretation of what a judgment is ...
Serenity
@qmikew1
Mar 31 2016 21:50
knowing the most difficult challenges is good - maybe like a study on how folks deal with those would be neat -- like there's always a very difficult 'something' in whatever that tests your ability to stick with something
Alice Jiang
@becausealice2
Mar 31 2016 21:50
@zcassini Filling in the blanks for someone is the worst form of teaching. If someone understands the concept and is just forgetting the syntax, it's life saving, believe me I am well aware. But coming into an IDE as a beginner and expecting to learn how to program from autocompletion is foolish. Don't think of FCC as a tool for people who are coding, it's a tool for people who are learning to code. BIG difference
Since when is judgement subjective?
evaristoc
@evaristoc
Mar 31 2016 21:51
agree with @zcassini: they are not getting that far yet... but there is a lot done in relation to some aspects of coding... AI is about that: I am not sure if still implemented in recent projects, but there are algorithms that create code
bitgrower
@bitgrower
Mar 31 2016 21:53

all human language is subjective, @alicejiang1 -- if that weren't true, there would be no need for lawyers and judges and what was "meant" .. our understanding of what a word means is based on our experiences in the world ... words are a summation of those experiences ...

put another way -- language does 3 things to our experiences ... deletes, distorts and generalizes ...

Alice Jiang
@becausealice2
Mar 31 2016 21:53
If that were true then there wouldn't be common languages at all :laughing:
bitgrower
@bitgrower
Mar 31 2016 21:54
I believe we understand FAR less of what is in other person's world than we think ... or, as one of my teachers once put it, if we were to experience the world as another person does, it would be a VERY altered state ...
Alice Jiang
@becausealice2
Mar 31 2016 21:55
language and experience are not synonymous
Brian
@sludge256
Mar 31 2016 21:56
There is a Clippy plugin for ReSharper btw.
clippy
evaristoc
@evaristoc
Mar 31 2016 21:56
@alicejiang1 I don't know... I agreed with you, but put yourself in the time when no calculators were available...
Those who support the use of techonology claim that it is to help you to reach faster the most complex problems above... If the computer can code loops for you, you dedicate your time to solve even more difficult problems...
Serenity
@qmikew1
Mar 31 2016 21:56
auto completion isn't necessarily bad (or linters etc.) - If somebody abandons their pursuit because they didn't have correct closing (insert whatever) stuff (but otherwise their logic was sound) - I think that would be sort of tragic (yes, it's implicit that they would discover this) -- just chiming in
bitgrower
@bitgrower
Mar 31 2016 21:56
again, you and I will disagree ... what else is language, if not a "handle" for a set of experiences ... those experiences can include sitting in a classroom, and hearing someone spout a set of words explaining what something else is ... but human perception is a pretty weird thing ...
evaristoc
@evaristoc
Mar 31 2016 21:56
I am not saying that this is good...
bitgrower
@bitgrower
Mar 31 2016 21:58
but, for purposes of this room ... I would suggest we just agree to disagree ...
Brian
@sludge256
Mar 31 2016 21:58
There's nothing wrong with that.
Serenity
@qmikew1
Mar 31 2016 21:58
that clippy
bitgrower
@bitgrower
Mar 31 2016 21:59
that clippy is pretty wild ... :)
Zac Cassini
@zcassini
Mar 31 2016 21:59
bitgrower
@bitgrower
Mar 31 2016 22:00
interesting ... dang, time to go to work .. will have to look at it later ... :)
evaristoc
@evaristoc
Mar 31 2016 22:00

I think this could be a fascinating discussion: Data Science is also about what Sciences will pose to us in our futures

Everyone who is in Data Science has attended to a conference about the effect of Big Data in our lives at least once.

The impacts could be huge
And they are just by the corner...
Serenity
@qmikew1
Mar 31 2016 22:02
and so misunderstood (just wanting to help and all) - I'm sure MS must have done some studies on how a paperclip (that was slightly invasive) or wizard could have elicited such feelings of anger - before they shipped this thing in the 2000s (like the group meeting somebody said "yeah this is a good idea") lol
evaristoc
@evaristoc
Mar 31 2016 22:03
@zcassini :+1:
Zac Cassini
@zcassini
Mar 31 2016 22:03
@evaristoc I thought you'd like it if you hadn't seen it before.
evaristoc
@evaristoc
Mar 31 2016 22:04
No, I haven't... really nice...
Serenity
@qmikew1
Mar 31 2016 22:04
ok, if software is checking software (auto bug checking) isn't this sort of skynet
Zac Cassini
@zcassini
Mar 31 2016 22:05
nah
bitgrower
@bitgrower
Mar 31 2016 22:05
@qmikew1 -- it's interesting how corporate decisions might be made ... including how much testing they did for unleashing it on the world ... one of my old bosses did consulting on Software QA for msft ... sometimes they throw stuff out there to see how it goes ... aka have the users "beta test" the software ...
Zac Cassini
@zcassini
Mar 31 2016 22:05
your virus scanner has been checking software for a long time
Serenity
@qmikew1
Mar 31 2016 22:05
different
it's not learning... it's being fed defs
Zac Cassini
@zcassini
Mar 31 2016 22:06
they don't all work on defs
Serenity
@qmikew1
Mar 31 2016 22:06
and I use linux
lol
true there is heuristics
Zac Cassini
@zcassini
Mar 31 2016 22:06
they can really do some complex analysis
bitgrower
@bitgrower
Mar 31 2016 22:06
that's actually the old stuff ... read up on Norton, which analyzes software behavior ...
Serenity
@qmikew1
Mar 31 2016 22:07
the test in the wild (i.e., decisions on what to release comment) on users is sad but true I think @bitgrower
Zac Cassini
@zcassini
Mar 31 2016 22:08
even compilers make complicated decisions on how to take waht you have written and how to write that.
bitgrower
@bitgrower
Mar 31 2016 22:08
hey ... the dwim instruction !!!
Serenity
@qmikew1
Mar 31 2016 22:08
norton is a bloated piece of cra^ that kills performance and hooks all kinds of stuff... the perceived gains of of not having a virus (for those that use) is because their machines don't work long enough for them to be exposed to anything
Zac Cassini
@zcassini
Mar 31 2016 22:09
yeah but norton was the shizzle back in the DOS days.
Serenity
@qmikew1
Mar 31 2016 22:09
lol
yes, there's that
evaristoc
@evaristoc
Mar 31 2016 22:10
When, sorry? DOS? What's that??
Joking...
I've got this book somewhere
evaristoc
@evaristoc
Mar 31 2016 22:12

@zcassini :

The system was tested against seven common open-source programs that had bugs, and it was able to repair the code in all instances between two to 10 minutes per patch, according to the research team behind it.

Zac Cassini
@zcassini
Mar 31 2016 22:12
@evaristoc Why won't they run it against firefox!
Serenity
@qmikew1
Mar 31 2016 22:14
all of this is comparison to known things.... the skynet part is when machines (kind of a leap from the checking for bugs (again, against known projects like in that example (a sample)) is when they begin to write viruses - although I don't think that the machine 'spark' in ideas will ever be comparable to people's ingenuity
bitgrower @bitgrower is not at all certain of that ...
Zac Cassini
@zcassini
Mar 31 2016 22:16
is that a transformers reference?
Serenity
@qmikew1
Mar 31 2016 22:16
no, that's terminator
lol
'skynet'
evaristoc
@evaristoc
Mar 31 2016 22:17
@zcassini Hahaha! Shhh... the firefox maintenance team is hiding the article out the view of the firefox CEO
Serenity
@qmikew1
Mar 31 2016 22:18
no, 'spark' is a general term that predates transformers (never heard of it in that context) for an idea
evaristoc
@evaristoc
Mar 31 2016 22:19
People I have to go! Late here... take care!
Zac Cassini
@zcassini
Mar 31 2016 22:19
@evaristoc cya
Serenity
@qmikew1
Mar 31 2016 22:19
later
Zac Cassini
@zcassini
Mar 31 2016 22:21
@qmikew1 You sir are correct. Robert Fludd, an Englishman born in 1574, spoke of the Cabalistic angel Metatron and the 'world soul', of which individual souls are sparks struck from it - p220-1 'The Age of Genius' A.C. Grayling.*
bitgrower
@bitgrower
Mar 31 2016 23:36

@bitgrower intrusive? not really... a bit artificial perhaps... Artificial Intelligence... it is interesting how fashionable such a non-human way of interaction could get...

Already there ... they've done interesting stuff using robots with elders ... and some reports have the elders actually preferring them ... google: robots with elders for a ton of hits ...