Data Science with Julia

This just came in the mail today! Looking forward to diving head first into this!

Shout out to the authors if they are on the forum

34 Likes

Hi,
What’s your impression, can you already say something?
Guenter

Looks pretty excited, I’d say :grin:

7 Likes

So far so good. There are no practice questions at the end of the chapters which is a bummer but the code examples are helpful.

1 Like

Thanks for your feedback. Are the code examples at least practical examples or dummy examples?

Definitely practical with sense of “this is how you should code this”

1 Like

What Julia version do they use?

It is 1.0 compliant

4 Likes

Glad to see Gadfly plots front and center :heart_eyes:

2 Likes

Upon further reading and completion of this book, it is not very good. But it has made me a better programmer having to work through all the poor coding examples. This book doesnt cover linear regression if that says anything…

Also slightly off topic - My boss wouldnt pay for JuliaCon this year, so in protest, i watched what i could during the livestream. I found the results of Julia survey enlightening but not surprising. Seems the community is suffering from lack of diversity. Well enter me, a black male, who really loves this community and loves the language! I’m hoping as the community grows, our community becomes more diverse.

I am wondering if there are any seasoned Julia data scientists looking to mentor me? I’m located in Boston(Roxbury) and work at non-profit community health center (<- redundant). I’ve been teaching myself Data Science for a couple of years (started with R) and often find myself being lost in a gigantic sea of information! I would definitely benefit from a mentor and our community becomes more diverse.

11 Likes

Maybe a formal mentorship program/interface would be worth thinking through at some point. It would ideally lower the barrier to proper mentorship (a time consuming and non trivial task that often requires training and curriculum go be performed properly)

1 Like

This is not surprising; data science is a vast and diverse field encompassing various methodologies, and a lot of practical skills. I would recommend that you focus on something concrete.

You could work through some online courses in areas that interest you, then ask questions on this forum when you need help.

You can find open course materials eg here:

https://ocw.mit.edu/courses/find-by-topic/#cat=engineering&subcat=computerscience&spec=datamining

https://ocw.mit.edu/courses/find-by-topic/#cat=mathematics&subcat=probabilityandstatistics

5 Likes

Yes, this is very true. It applies when entering almost any field. Here is something that may inspire the OP, given his current working background:

The Rotterdam Study (official website, wiki)

It is a well-known example of an early and ongoing application of data science in public health care. (I am not involved in this project, but live and work nearby.)

I much commend your enthusiasm and - having looked at the website of your current employer - also much commend their goals and purpose. There are two remarks I would like to make:

  1. Personally, I really don’t care whether somebody is black, white, yellow… What matters for me is the combination of knowledge, intention and working spirit. Of course I cannot be familiar with your background knowledge, but I see that in any case you got a lot of the latter two.

  2. I am a mathematician (not a data scientist), so it may be my own inclination, but I would focus first and foremost on strong fundamentals (in applied statistics and data science, in your case) in a somewhat language-agnostic way. That is, I would perhaps start with reading material that does not rely too much on one particular language, as such material tends to be outdated quite quickly. If you then combine what you read with self-assigned exercises and experiments (for this Julia seems very good), then you have the best of both worlds: A solid language-independent basis and hands-on experience in a modern, actively developed language.

If you are interested in the Rotterdam Study, I could try to get you in touch with people from the university here, but I stress that I cannot make any promises, as I am not involved in the project.

Best of luck with your endeavors!

6 Likes

Thank you for your reply! Extremely helpful. Do you have any recommendations for reading materials? It takes a lot of courage to ask for help so I appreciate the kind words and non presumptuous response.

1 Like

I would recommend two textbooks on practical statistics first.

The first is the famous “Data Analysis Using Regression and Multilevel/Hierarchical Models” by Gelman and Hill: http://www.stat.columbia.edu/~gelman/arm/ The book has practically no theorems or proofs, but it does have a lot of models and examples, many of them from the life sciences. Working through this book with modern tools is a real education.

The second is “Statistical Rethinking” by McElreath: https://xcelab.net/rm/statistical-rethinking/ which is an eye-opener about modeling in practice, especially about model validation and what can go wrong. @goedman has programmed most (all?) of it in Julia,

using various packages. This can also serve as a Rosetta stone into the relevant package ecosystem.

16 Likes

Please also have a look at my co-authored “Statistics with Julia (DRAFT)”: https://people.smp.uq.edu.au/YoniNazarathy/julia-stats/StatisticsWithJulia.pdf

Would love to hear what you found helpful and what not. Regarding mentoring, happy to discuss more - feel free to contact me. Yoni.

9 Likes

The current set of packages in StatisticalRethinkingJulia is an ok starting point to ‘take’ the StatisticalRethinking course. All important models are demonstrated in Stan, DynamicHMC and Turing in the included …Models packages.

The current version (0.9.0) is basically a first time around. Version 1.0, which I expect to release later this year, will be a significant upgrade as indicated in the README of StatisticalRethinking.jl and is linked to updates to other packages (StanJulia, DynamicHMC and Turing/AdvancedHMC).

7 Likes

I did it! I did it! I am finished with my book “Data analysis with Julia”. It is written in German language and I hope that Julia will be used in Germany as a tool for data analysis! Coming from the statistics software R, the introduction was a challenge from time to time. But thanks to this community I succeeded well! For this I thank the Julia community and am happy to be a part of it!

Amazon-Link

19 Likes

Congrats! You should start a new thread with the announcement :slight_smile:

1 Like

The list of non programming languages I need to learn just keeps growing. Are there plans to translate it at all?

1 Like