These are chat archives for FreeCodeCamp/DataScience

18th
Dec 2017
Josh Goldberg
@GoldbergData
Dec 18 2017 02:54 UTC
Eh. No prob! Honest typo! I thought it wasn’t a typo until after I read the article 😊 @erictleung
Alice Jiang
@becausealice2
Dec 18 2017 04:18 UTC
is "One of the first lessons I learned starting in data science is that size does, in fact, matter." an awful first ~100 characters for a linkedin summary? (That's about how much you have before the "see more" on linked in mobile
I should rephrase my question, would you click "see more"? The rest is going to be all the professional blah blah blah about distributed computation and "big data" :/
Eric Leung
@erictleung
Dec 18 2017 21:02 UTC
@becausealice2 haha nice :smile: I think it depends on the culture at whatever company you're going for. I hear that working well with other co-workers is important for productivity so you might have a chance with that headline. I definitely get the impression that you have a sense of humor when reading it. Hopefully recruiters or HR peeps will be interested as well.
Ran into this summary article for a large report put out by JP Morgan on machine learning and AI. It picks out some tables listing software packages in Java, Python, and R they think will be important for machine learning (they did leave out data cleaning and manipulation packages though e.g. Python's pandas). One point from the article was "You won’t need to be a machine learning expert, you will need to be an excellent quant and an excellent programmer" which goes to some discussions we all had earlier on focusing on results rather than advanced new methods. Here's a link for the full 280 page JP Morgan report in case you're interested.
Josh Goldberg
@GoldbergData
Dec 18 2017 21:11 UTC
@erictleung you’ve been sharing some gold lately man. Thanks.
CamperBot
@camperbot
Dec 18 2017 21:11 UTC
goldbergdata sends brownie points to @erictleung :sparkles: :thumbsup: :sparkles:
:cookie: 560 | @erictleung |http://www.freecodecamp.org/erictleung
Josh Goldberg
@GoldbergData
Dec 18 2017 21:11 UTC
I may get a masters in CS for fun
Eric Leung
@erictleung
Dec 18 2017 21:25 UTC
@GoldbergData anytime! Just sharing things I've found interesting :smile: Glad some people are finding some use in it haha. Do you need a masters degree? These days, it seems you can get away without doing a masters. I do have to say it will give you some structure in designing a programming project.
Josh Goldberg
@GoldbergData
Dec 18 2017 21:39 UTC
I don’t need. But if my job assists in paying for it why not? What’s more valuable, dual masters (analytics and CS) or one PhD (stats, CS?, AI?)
I’m in a position where I can pursue a couple masters for the next 5 years, or do one masters then a PhD. The former allows me to work all the way through. The latter I’d have to quit my job at the PhD level. Are either necessary? Maybe, maybe not. What’s the best route? I’m not sure.
@erictleung
Alice Jiang
@becausealice2
Dec 18 2017 22:04 UTC
This article passed my Twitter timeline and thought I'd share
Very important overall message about data quality being fed into ML models
@erictleung Glad you like it ^^
I asked in fcc/fcc since that chat is more active and got a lecture on professionalism that I had to finally just ignore because... well... reasons :joy:
Alice Jiang
@becausealice2
Dec 18 2017 22:10 UTC
@GoldbergData I'm formal-education-averse, so definitely not the one to ask, but I imagine a lot of the decision process should surround your end goal, and even then it won't hold you back if you seek work in some other field, depending on said field.
Eric Leung
@erictleung
Dec 18 2017 22:12 UTC
@GoldbergData that's a sweet deal if your job assist with paying for the masters :+1:
@becausealice2 nice article :+1: I think I passed by that conversation on Twitter. It does highlight the issues of doing machine learning in healthcare. The data are messy and will take time to get that data in good shape.
Alice Jiang
@becausealice2
Dec 18 2017 22:15 UTC
DS/ML in healthcare scares me
I would be sad if I screwed up and people lost money or whatever, but healthcare is much higher stakes :(
Eric Leung
@erictleung
Dec 18 2017 22:16 UTC

Compared to human visual assessment, the labels in the ChestXray14 dataset are inaccurate, unclear, and often describe medically unimportant findings.

:worried:

Josh Goldberg
@GoldbergData
Dec 18 2017 22:16 UTC
@becausealice2 @erictleung interesting thing is I continue to informally educate myself despite being enrolled in a formal program. I do this stuff for fun. Not because school told me to. Why not formalize my passion if the cost is low? (In the case of the dual masters). The big question is how (and if) a PhD would help me professionally and intellectually? I guess it depends if I want to do AI research. It seems one of the obvious benefits of a PhD is if you want to do research. Otherwise, staying in industry, I’m not sure how fruitful a PhD would be outside of already having a masters
Eric Leung
@erictleung
Dec 18 2017 22:17 UTC

@becausealice2

I would be sad if I screwed up and people lost money or whatever, but healthcare is much higher stakes :worried:

Oh yeah, that can be a big issue. However, the responsibility should always be in the hands of the physician. I don't believe that computers should have the end all say all on medical decisions. So maybe you won't have to worry as much.

Alice Jiang
@becausealice2
Dec 18 2017 22:18 UTC
Excellent point.
This table at starbucks... The pole that holds the tabletop up is not secure in either the base or the top itself so I have to type very gingerly or it starts shaking and swiveling :/
Eric Leung
@erictleung
Dec 18 2017 22:19 UTC
@GoldbergData I think the benefits of having a formal training is really about the environment and community of students and professors you'll have to learn from. Otherwise, if you have that community and expertise you're able to leverage, then the formal masters might not be a great benefit to you. My two cents.
@becausealice2 oh those are the worst... :-1:
Alice Jiang
@becausealice2
Dec 18 2017 22:23 UTC
I am having trouble figuring what else to say in my summary lmao
Alice Jiang
@becausealice2
Dec 18 2017 23:06 UTC
I'm gonna barf. I just looked up linkedin's top big data developer profiles and it gave me results around the Greater St. Louis area and most of them work, or have worked, for Monsanto
Josh Goldberg
@GoldbergData
Dec 18 2017 23:32 UTC
Lol
Alice Jiang
@becausealice2
Dec 18 2017 23:44 UTC
For anyone interested, Riot Games senior data scientists did an interview at Spark Summit 2017 on how they do their DS thing