data-science

Data science, data analysis, and machine learning in Clojure https://scicloj.github.io/pages/chat_streams/ for additional discussions
Aviv Kotek 2020-04-18T13:18:09.101400Z

matplotlib equivalent in clj?  `vega/oz` would be enough? is there something else I should look on?  I see some cluttered reddit threads but not much more, ty

Aviv Kotek 2020-04-19T07:01:21.105900Z

curious if using libpython-clj and starting to embrace the interop would be a good way to deal with those stuff

Aviv Kotek 2020-04-19T07:01:24.106100Z

will give it a look

2020-04-19T09:38:30.108200Z

Is there anything from vega or plotly that do not cover your need with matplotlib? I find the data orientation of these library better than calling functions on mutable objects.

Aviv Kotek 2020-04-19T11:49:58.108500Z

I still haven;t looked deeply into vega, but generally I would like to have tools which are similar to matplotlib (without the need to learn a new 'language' over again on every data task in clojure)

Aviv Kotek 2020-04-19T11:50:42.108700Z

i'll give it a look this week and let you know if I see anything. but for example yesterday I spent half a day trying to do a simple plot/scatter of words in 2d. then I used clj-xchart but it doesn't support text/tagging on the points themselves 😣

daveliepmann 2020-04-19T16:34:17.110200Z

I'm not sure what you mean by 'new language on every data task' — Vega/Lite is a DSL, but a fairly thin one and it covers most data visualization.

Aviv Kotek 2020-04-19T20:09:40.114800Z

yep, was upset with the need of a new DSL, by the way - it's awesome. the data concept is great. thanks

genmeblog 2020-04-18T17:48:12.101500Z

I'm writing (and rewriting now) cljplot. Also if you are not afraid interoperability take a look at clojisr and use ggplot2. Or call matplotlib via libpython-clj.

daveliepmann 2020-04-18T18:04:18.101700Z

Incanter has a bunch of presets. Because Clojure is a relatively small community compared to e.g. JavaScript/JVM/Python, for this kind of thing (general-purpose problem that lends itself to a DSL with lots and lots of features and therefore bugs) I like piggybacking on solutions that aren't limited to our world. So I went with Vega (99% Vega-Lite), but other choices make sense too.

2🚀
practicalli-john 2020-04-18T18:17:43.105100Z

Practicalli broadcasts on getting started with data science visualization. I cover some basics of data extraction and transformation (clojure.data.csv, semantic-csv, metosin/jasonista), repl visualisation with ascii-graph and visualization with Oz (vega-lite). https://www.youtube.com/playlist?list=PLpr9V-R8ZxiDUXIR2z8Y8wvhpoPyl0t_D

7👍