These are chat archives for FreeCodeCamp/DataScience

3rd
Jul 2016
Alice Jiang
@becausealice2
Jul 03 2016 02:11
JS is barely even a real language and not worth my time or else I'd have stuck with it I'm the first place. If I had a choice in this project, I wouldn't have even considered JS.
Sam Aiken
@SamAI-Software
Jul 03 2016 02:19
@alicejiang1 what language would you consider then?
Alice Jiang
@becausealice2
Jul 03 2016 02:22
Usual suspects... R or Python.
They could actually do the work well and without all the headache, and I know enough of either that I probably could have had everything long since done by now.
Sam Aiken
@SamAI-Software
Jul 03 2016 02:25
:+1:
evaristoc
@evaristoc
Jul 03 2016 09:36

Yesterday I went to a Hackathon for Data Scientist about Endangered Species in Africa. Shame on me, I brought the wrong hardware/software: a Samsung Notebook I have, with the minimal implementation, just python3 (that's all!). That was unfortunate because I couldn't help that much my team, where we focused on species displacement and migratory patterns along time. However I at least made a d3 map plotting the migration path of the species under concern - after going through a lot of debugging and computer crashing, of course...

We lost to very good teams and individuals (actually, I think we lost to everyone...): the winners presented even a simulation in GoogleEarths and had something already prepared for the hackathon; someone made a model of speed of movement of the different species; there was another team presenting relevant data about daily activity of some other species; but perhaps the best one in my opinion was someone presenting a predictive model of positioning (which was very much of interest for the hackathon objective).

The good news is that without knowing d3, I wouldn't have been able to make even a small programatic contribution that afternoon ( I help mostly discussing the results). Considering that I came ill-prepared to the hackathon, that is not too bad...

evaristoc
@evaristoc
Jul 03 2016 10:13
I must add that I didn't have real plans to participate: just to see how the process went; but eventually I felt I had to join a team, otherwise I would have been doing nothing there...
Sam Aiken
@SamAI-Software
Jul 03 2016 10:32

(actually, I think we lost to everyone...)

:smile: :smile: :+1:

evaristoc
@evaristoc
Jul 03 2016 11:23

@SamAI-Software
:) :) I was not really planning to work with my computer: the event was expecting around 30 participants: at the end only 10 showed off... One of the teams was already a partner of the project and they were already well prepared... it was a very nice project! In general, all projects were really good! And the topic... Excellent!

I am going to do this more frequently...

Sam Aiken
@SamAI-Software
Jul 03 2016 11:24
I see, no problem, just good sentence, make me laugh :)
evaristoc
@evaristoc
Jul 03 2016 13:54

@alicejiang1 @samosale there is a first file with facebook data at this gist if you are interested. I managed to update 600 records and added a list of events (which would include past events too...).

I am still running into some trouble with fb. One thing that I just discovered is that after making the request through my Firefox to the update page, it goes about 30 seconds not having receiving notification while node is still downloading data, and then Firefox ping my server for a second answer before accepting a no response. Although that doesn't prevent node to continue searching, the ping start a new search which duplicate the number of requests on fb and therefore exhausting my rate limit. @samosale: any idea how to control for the browser not to ping the server?

evaristoc
@evaristoc
Jul 03 2016 18:47
@alicejiang1 R or python: I would say --- try R if you are going for statistics, python if you think you can be more into programming... R has a longer learning curve for the language but if you know statistics it will be more productive; python is the other way around: it might looks easy at first, but if you are not confident with statistics you would find harder to use it for statistical analyses...
Alice Jiang
@becausealice2
Jul 03 2016 19:11
I'm no master, but I have a good enough grasp of statistics that R hasn't been that rough so far. That comment was regarding the campsite map specifically, though. JS is not the language for this kind of work.
Aleksandar B.
@aleksandar-b
Jul 03 2016 20:25
@evaristoc sorry, I can't say nor understand quite well your question, I will have to look some code examples... But if I understood correctly you are fetching FaceBook Api via browser? Why would you do that?
Alice Jiang
@becausealice2
Jul 03 2016 20:34
@samosale he's calling from the server.
It's set up so the server runs after a request from the browser, and the information is returned to the browser. He seems to be having a problem with Firefox re-sending the request before the server is done and the whole process resets and starts over.
I think...
Aleksandar B.
@aleksandar-b
Jul 03 2016 21:21
@alicejiang1 ah, ok
In that case i need to look at code
to figure out what is wrong
Alice Jiang
@becausealice2
Jul 03 2016 21:23
I'm not sure how up to date he's keeping the repo on github, but it's the development branch on his fork of my FccCityScraper repo @samosale
evaristoc
@evaristoc
Jul 03 2016 22:03
@samosale @alicejiang1 thanks both of you!
I solved, @samosale:
  • to trigger the function in node, I called from an endpoint (/googlelists)
  • I had a function that closed the request at the end of a long cycle; meanwhile the browser was waiting
  • didn't know that the browser if not receiving any confirmation before certain time would ping the server automatically and if not response after the ping, it would close with a "no response"
  • the problem was the ping: it was a SECOND request that mounted on top of the first request, duplicating the calls to FB
  • solution: use only one request and close the request immediately, no point to let the browser hanging waiting for a response from server: the server would find the data anyway and save it into a file
  • after that, a problem with Exceding the limits of request to FB is also apparently gone
CamperBot
@camperbot
Jul 03 2016 22:03
evaristoc sends brownie points to @samosale and @alicejiang1 :sparkles: :thumbsup: :sparkles:
:warning: could not find receiver for alicejiang1
:cookie: 643 | @samosale |http://www.freecodecamp.com/samosale
Alice Jiang
@becausealice2
Jul 03 2016 22:07
I just found Google Feud.com , which is like Family Feud, but with Google's most common autocompletes instead of an audience survey. :D
Yay DS is fun :clap: :D
That should be www.googlefeud.com but I can't go back and fix it on mobile, so...
evaristoc
@evaristoc
Jul 03 2016 22:24
@alicejiang1 fun :)