data-science

Data science, data analysis, and machine learning in Clojure https://scicloj.github.io/pages/chat_streams/ for additional discussions
genmeblog 2020-04-23T13:04:26.131400Z

http://tech.ml.dataset compared to R data.table. All stuff works (with some additional helpers). See it here: https://github.com/genmeblog/techtest/blob/master/src/techtest/core.clj

➕ 4
🦜 3
2020-04-23T14:37:45.132300Z

anyone here got a recommendation for a good linux (ubuntu) friendly workstation? I think I'm going to be at home for a while and it is roughly upgrade time for me.

2020-04-24T13:33:21.136700Z

I was thinking desktop. I've been pretty happy with Dell laptops but was wondering if I could get more quack for my quid with a desktop

practicalli-john 2020-04-24T14:51:40.136900Z

@otfrom take a look at the Thinkpad P series desktops, the P330 and P330 Tiny can be configured as performant as you want (can afford). I was very tempted to get the P330 Tiny, its amazing what you can fit in there, but would cost a little more than the P330 standard size desktop https://www.lenovo.com/gb/en/workstations/#viewprodfragment

2020-04-25T09:17:39.137300Z

Thx

jsa-aerial 2020-04-23T14:40:04.132500Z

Would be even nicer if the actual R code was given. I know - easy for me to say 🙂

genmeblog 2020-04-23T14:41:19.132700Z

Why do you need R code when you have clojisr 🙂

jsa-aerial 2020-04-23T14:42:41.132900Z

It would give a 'better' comparison for those who use R

genmeblog 2020-04-23T14:51:55.133100Z

no problem then, I will add it soon.

1
genmeblog 2020-04-23T15:09:33.133400Z

done! 🙂

👍 1
2020-04-23T17:11:04.134100Z

take a look at http://system76.com ?

2020-04-23T17:15:13.134400Z

Is there identical performance?

genmeblog 2020-04-23T17:16:41.134600Z

No. Reading table from net Clojure is 4x slower. Filtering 10+ times slower.

genmeblog 2020-04-23T17:16:55.134800Z

Chris is investigating this now.

🆒 1
2020-04-23T17:20:41.135Z

data.table is mighty fast, so approx performance would be incredible

2020-04-23T17:21:18.135200Z

my understanding is that a lot of work went into making fread so good

genmeblog 2020-04-23T17:25:50.135700Z

Also it is C based. Anyway my experience with tech.ml.database is good. flights has 250k rows.

2020-04-23T17:36:19.135900Z

Desktop or laptop? I've had a pretty good time with my Lenovo Carbon X1. The Dell XPS series look pretty tempting too though, and some of them you can actually get to ship with Ubuntu (or at least used to be able to).

2020-04-23T17:37:03.136100Z

For laptops, none of the system76 seemed to strike the right balance to me for power + portability. I have one of their desktops though, from years ago now, and it served me pretty well.

practicalli-john 2020-04-23T19:12:52.136300Z

I am loving my 2nd Gen Lenovo extreme workstation, it has been well worth the investment. It feels like I can run all ther repl's in the world. https://www.lenovo.com/gb/en/laptops/thinkpad/x-series/X1-Extreme-Gen-2/p/22TP2TXX1E2