search

ledell / useR-machine-learning-tutorial

useR! 2016 Tutorial: Machine Learning Algorithmic Deep Dive http://user2016.org/tutorials/10.html
Apache License 2.0
402 stars 206 forks source link

Convert tutorial into Bookdown #7

Open ledell opened 6 years ago

ledell commented 6 years ago

I am looking for a volunteer to convert this tutorial into Bookdown format so that the content is easier to navigate. Once it's in Bookdown, we can find a place to host it on the web.

If you are interested, please let me know! (please reply here or send me an email)

bradleyboehmke commented 6 years ago

@ledell ,

@bgreenwell and I are working on an R book with CRC press and have started building the Bookdown format. In fact, our book will hit on some of the same topics as it has a big emphasis on predictive analytic techniques with R. We don't think it would be that big of a challenge to convert your current material to Bookdown format (and extend any parts of it if you desire). We're up for the challenge if you'd like to partner.

ledell commented 6 years ago

@bradleyboehmke @bgreenwall That sounds great!

So the main thing to decide is whether we:

  1. Keep the Rmd files in this repo as the main source of the text (and maybe remove the ipynb files if we convert the missing two chapters to Rmd) so that there's only one source. When I created this tutorial, the R kernel for Jupyter had just come out and I thought it was cool so I used that instead of Rmd (the Rmd files actually came from an external PR), but now I think Rmd is clearly the preferred format and so I'm okay with removing the ipynb notebooks.
  2. Keep the code in here as-is, copy the Rmd files over to a new repo and have that be a "live" version of the text which could possibly grow and be extended over time. If there is no immediate plan to extend the content, then this might not be necessary. Personally, I don't have plans to extend this right now, but perhaps I would at some point in the future. I am open to contributions though.

I guess the plan when it's done would be to host it on Github pages.

bradleyboehmke commented 6 years ago

We're open to either approach but our initial thought was #2 - probably easiest to leave these source files as-is and start a new repo by copying one .Rmd over at a time. I could see @bgreenwell and I contributing some new material over the next year as well.

And yes, host on Github pages and push to https://bookdown.org for sure.

We'll wait for confirmation but if you're cool with #2 then we can start pressing. Just let us know.

ledell commented 6 years ago

Hi @bradleyboehmke sorry for the delay... this got buried in my notifications. It might make sense to create a github org to host & collaborate on the new repo. What do you think?

bradleyboehmke commented 6 years ago

No problem @ledell . We've got the scripts for each chapter complete and I was able to build the bookdown locally but @bgreenwell couldn't quite reproduce on his end. Probably just a minor error we need to debug but hopefully we'll have a fully operating bookdown version next week for your review.

Right now we're just working out of a repo in a github org that @bgreenwell I maintain but when its time to publish/host it wherever you'd like...we're flexible! For now, I just added you as a collaborator.

bradleyboehmke commented 6 years ago

@ledell I was able to get the bookdown version fully operational. You can check out the current version at https://koalaverse.github.io/machine-learning-in-R/. This is just a holding place for now as we edit and get it finalized. We can move it somewhere else when we're ready to permanently publish it.

@bgreenwell and I will start to do minor editing. When you get a chance could you peruse it and let me know if there are any edits/additions you see are necessary?

ledell commented 6 years ago

@bradleyboehmke Thank you! It's great to see this in bookdown format. :-) I have some major changes to make to the Stacking code section (I'll make a branch and work on that), but everything else should still be reasonably current. Let me know if you see any out of date topics or code and we can discuss how to improve them.