Clojurians Log v2
data-science 2024-04-01
Clojure camp has very a nice map for "what's worth learning" if one wants to learn Clojure: https://currmap.clojure.camp/ Anyone know if there's a similar map for "what's worth learning" if one wants to learn to do data science with clojure?
Great idea. Let us make it?
I'm in!
Though I'm not sure what should be on it.
Or who we should ask to figure out what should be on it. But you might have an idea about that?
Let us think. Here are a few ways we could create such a table: • systematically go through a textbook that we care about (or some part of it) to collect topics • run a survey and ask what people wish to know about • keep it an evolving structure, and update it by following community activity for some time (chat questions, blog posts, etc.)
What do you think?
I’d probably try to get a quick and dirty first draft up first, then ask people for feedback. Ideally, people who both know the data science, and like to teach. We could show the quite nice table for all of clojure, and compare to our very early draft for the Clojure data science table. --- That’s what I’d do to start. After we have something, I think both going through a textbook, running a survey and following community activity would be great.
I think you're right.
I can create a draft, then we can look at it later?
That would be great. This week I'm a bit committed to other tasks, next week I'd be glad to join you with that.
Cool. I’ll make something and give you a link when I’ve got something to show 🙂
A related question, that you may have some thoughts about, is how to index existing tutorials (even if they currently do not cover the topics which are actually needed). Just started a Zulip discussion of it with a few of the relevant people: https://clojurians.zulipchat.com/#narrow/stream/151924-data-science/topic/scrapbook
Zulip link appreciated, I don’t follow everything that happens there!
how to index existing tutorials (even if they currently do not cover the topics which are actually needed).I’d put EDN files in a github repository, write code to create a datascript / datomic DB from those files, then generate a variety of HTML tables. One HTML table per “kind of index” we care about. Sort of like the tables I made here: https://play.teod.eu/clojure-europe-easter-2024-reading-recommendations/ (though no code and no datascript was involved in the making of those tables, that was just manual work.)
Nice
Very nice
Here’s a very early draft for curriculum map for data science: https://play.teod.eu/currmap-clojure-data-science-draft/currmap I took several shortcuts. • I didn’t make the nice, hierarchical table layout on the left that Rafael did for his table, I just used / to indicate nesting. • I copied in sample data from his table to get started with this table. • I have not put in much data, yet. The code involved is two files. An https://github.com/teodorlu/play.teod.eu/tree/eb03fc2cd5356a2947f9b754d659229789ba28e5/currmap-clojure-data-science-draft/currmap.edn, and a https://github.com/teodorlu/play.teod.eu/tree/eb03fc2cd5356a2947f9b754d659229789ba28e5/currmap-clojure-data-science-draft/render_currmap.clj to generate the HTML table.
To fill the table, I’d want to talk to people who are better at data science than I am. And work together either in Miro, or a collaborative spreadsheet (google sheets). After a “what’s worth learning?“-session, I’d translate to EDN (either by hand or write a program to read the collaborative spreadsheet).
Great