These are chat archives for FreeCodeCamp/DataScience
discussion on how we can use statistical methods to measure and improve the efficacy of http://freeCodeCamp.com
@bigyankarki working on a project about students (mainly fCC) who got jobs as web dev when no having previous coding experience. No big data, not even ML. Discourse analysis mostly, helped by some semi-automated procedures. Outputs are a medium article and an website profiling the sampled students. Most of the data is already collected, I am working on the article and the website at the moment.
By the way, @bigyankarki: checked your site: nice photos those you made!
@becausealice2 @GolbergData - agree with @becausealice2. It seems somehow to add to the decrease in trust after the failure of the data analytics companies during the US presidency forecast. The trust of the general public is likely eroded. Not only there results cannot be trusted: they are prone to manipulative practices using people's data. And I am sure CA is just an "unfortunate" one: there are surely several successful companies out there following similar practices.
It's easy to be proactive when thinking about things like physical possessions, but you simply can't know who has your information because it's not as simple as arriving home to a busted in door.
Then we as scientists of the data have accepted the responsibility of protecting that information on others' behalf, then here some money-hungry bastards come along with no ethics and ruin it for the rest of us
:) :) :)
From the texts written by 64 users from the full sample (medium or forum) I extracted some elemental topics I thought relevant for new coders.
I counted the times those writers mentioned those topics, at least once.
The most important aspect they wanted to talk about? Based on my classification (important!) "portfolio" and "commitment"were probably the most mentioned topics.
I showed this to an "expert" (aka a moderator with experience in the chatroom) and found some sense in the data.
Again, this is a categorization I made on my own and it is not validated. I tried to do my best to keep objectivity but this kind of work is usually very subjective. I hope to have some trust from you.
My purpose is to avoid any statistics discussion. Might not be applicable in this case. But not because it is not statistically proved doesn't mean it lacks relevance. It rather depends on what you say about what misleads.
One important thing to mention is:
I have no records of researchers doing this kind of characterization at least over this specific type of users (people who succeeded to find jobs as web dev without no previous coding experience). So here you go...