Any data scientists around? I'm about 5 months in to my self-education in the field. I have a BA in computer science and I'm trying to augment my skill set so I can get away from full stack web dev and do more back end ML stuff.
I took the Andrew NG ML class on coursera and loved it. If anyone in the field is around, I'd love to talk about what I should study next.
Are you struggling with the maths? Everytime I see someone from CS talking about datascience they all go about programming languages and the latest frameworks to do datascience but they seem to utterly ignore statistics and optimization
Joseph Scott
No, I'm quite comfortable with maths. I've been studying convex optimization and probabilistic graphical models for the mathematical underside.
Christian Ward
Anyone else studying Data Science? If so, what are you studying currently?
Levi Scott
Yes, you could self-educate yourself statistics and algorithms, but I doubt you'll be hired as a data scientist unless you have an actual degree (MSc) in either math/statistics or machine learning.
Evan Garcia
you sound like you're good to enter the workforce. Maybe learn how to distribute databases and queries, and how to structure your work pipeline. tech.marksblogg.com/ I find these blog posts to be of really good quality. Most big business have or need analytics support and analytics can increase sales consistently if you do it right. With most small-medium businesses this would mean not only doing the exploratory analytics to find where you can optimize the sales process, but also be able to quickly hack together something that works to optimize it and measure that.
Ethan Long
I started an LLC with a friend. We were thinking of doing consulting in the field first. Like finding small businesses with analytical needs, providing a solution, then using those projects as a source of credibility for larger projects.
Adrian Sanders
Honestly getting a nice pipeline in place is really what I'm struggling with. I know TensorFlow pretty well, but I don't have a way of storing trained models, versioning them, and providing front ends. What do you suggest in terms of a pipeline?
Ayden Watson
literal meme major
Evan King
I'm starting a two years M.Sc. Program in Data Science next month (got my BA in CS too). Got into it by some kaggle competitions and blogs.
Tbh, I don't really want to work in typical business domains like consulting, business analytics or marketing. I'd rather try get into pharma-/biotech or health care projects.
Kayden Diaz
>data science
I hate that phase so much. Just call it statistics.
Joseph Taylor
I'm tired of the phrase too, but data science is not simply statistics.
Data science implies statistics and modern methods like machine learning, neural networks, i.e. stuff that's a bit outside the boundary of stats.
Gavin Thomas
>machine learning
Is regression and classification AKA statistics.
>neural networks
Literally just using the chain rule. Are you seriously suggesting we need an entire new discipline created just because you're doing statistics + writing programs to solve basic calculus optimization problems by repeatedly finding the squared error of a known training set's answer's minus your network's answers and adjusting weighted connections based on their gradients with respect to the activation function you're using?
Ayden Morris
You're absolutely on point with your response. Data Science isn't really it's own thing. It more implies the combination of statistics and more computer science type stuff.
I've heard similar arguments leveraged against calling Computer Science, Computer Science instead of mathematics.
Really it's pedantic to argue the purity of terms in my opinion. People use the word to describe a particular application of statistics, and it's very clear what is meant when we say data science.
Nicholas Nguyen
I've also heard people argue that everything is simply philosophy, but it should be clear why we use other words to describe subsets.
Connor Long
Not really a data scientist since I have no Idea about programming outside of R, but I am a survey statistician. If you have questions I might be able to help
Owen Nguyen
Not him, but what was your degree in? I'm a bio/math double major with a statistics minor, but have no idea what I want to do yet upon graduation.
Isaac Sullivan
I actually have a bachelors degree in sociology (cause I was pretty fucking stupid), took lot's of survey methodology classes, learned sample designs etc. and did a Msc in survey statistics.
It's pretty nice actually I design and conduct and analyze surveys for a big company, focus on analyzing tho.
Christian Garcia
Not that guy, but I just want to say that a bio/math double major with stats will land you any cushy bioinformatics job you could ever want. If you know how to code or some computer science type stuff, you're especially in the gold because that's one of the hottest fields to be going into right now.
Luke Hernandez
NYU has a Center for Data Science lead by Yaun Lecun. I think arguing against the term is hipster.
Michael Bell
What statistical concepts have you found most useful in your line of work? What do you do on a daily basis?
Eli Butler
>focus on analyzing tho sorry for the hijack, but is this common for individuals hired with msc in stats to do more analysis than ground work?
Michael Hill
>applies for job >"Hi Mr user, have you got 5 years experience in the field X?" >"No? Hum, you are not exactly who we are looking for. Good luck"
Sebastian Miller
I'm in an introductory course, we have covered PCA and association rules so far. I think we are going cover decission trees now
Landon Hill
If you have to present results to your superior, there is not a lot of stuff you can do. I usually just put some graphs up cause they barely know what a t-test is (yes this is common). Also very important are weighting techniques, linear and multiple regression. And you better fucking understand sampling theory and how to manage complex sample designs if you want to do a good job.
It's my first job so I don't really know but I guess not.