Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Activity
    Jez Cope
    @jezcope
    I'm pretty conflicted about Jupyter vs Spyder — they both have pros and cons
    Tim Dennis
    @jt14den
    @jezcope Jupyter is v good for novices, removes features that IDEs have but that lets learners focus on Python
    Also from an instructor POV swc and data carpentry use Jupyter, makes it easier for mixing and matching and instruction prep
    Jez Cope
    @jezcope
    @jt14den yup, i think all the other arguments are so equally balanced that "what swc and dc use" is what wins!
    for the record i completely agree with jupyter reducing cognitive load, and the counterargument for using spyder is that it introduces both the interactive console and scripts (which are the most common ways to use python for batch work) without the additional load of manipulating a separate commandline session and text editor
    i'm happy to have the debate if we need it, but i suspect it would just cover the same ground that swc & dc have already covered so we might as well take the short cut and use jupyter
    Tim Dennis
    @jt14den
    The cognitive load argument always wins for me pretty much, I’ve taught SWC juypter where we end up making a module via the built-in text editor, so it’s not like we are limited from covering how to make modules in jupyter
    Jez Cope
    @jezcope
    ooh, that's interesting. do you just export it to .py from jupyter? i hadn't really thought of that workflow but it makes complete sense, especially for learners
    Tim Dennis
    @jt14den
    As I recall we copied them to a file from jupyter b/c we had a long notebook with a bunch of stuff. Then we opened a new notebook, imported the module and ran the code from the new notebook. There’s a lesson somewhere - think it was ‘intermmediate python’. Your suggested approach would be cleaner, switch to a new notebook when you start functions, get them working and ouput to .py
    Jez Cope
    @jezcope
    i like that approach
    depending on how advanced your workshop was you could have a cell at the end with unit tests in that you keep returning to and running; probably not suitable for lc python intro though
    i'd be interested in thoughts about how we could get test-driven development in as a natural thing without too much cognitive load though
    Tim Dennis
    @jt14den
    I’ve never taught the unit testing stuff in SWC py, have never gotten there, would enjoy doing it
    Maybe you could have as a demostration and then come back to it later?
    But seems like it would need to be intermmediary level
    Jez Cope
    @jezcope
    Taught an extended version of the OpenRefine lesson yesterday and it went down really well
    We got through all the material in the lesson, which took all morning and half the afternoon, and then had a "masterclass" session where we worked together as a class to figure out how to solve some data problems that some of the learners had brought with them
    I'm getting consistent feedback on the OR lesson that it needs more exercises though, and I agree
    Tim Dennis
    @jt14den
    @jezcope very cool! Did the learners let you keep their examples/dirty data?
    Jez Cope
    @jezcope
    Yes, and I'm going to get permission from them to incorporate in the lesson
    There shouldn't be any problem I'm told, because it's all library holdings info that's already publicly available in some form through our discovery portal
    Tim Dennis
    @jt14den
    That’s great, I’d like some actual numeric Data in the lesson
    Belinda Weaver
    @weaverbel
    Pasting this here since a bit of python chat is being had - sample lesson https://reproducible-science-curriculum.github.io/data-exploration-RR-Jupyter/
    Belinda Weaver
    @weaverbel
    Might be a good read for anyone thinking about creating lessons: https://carpentries.org/blog/2018/06/lesson-infrastructure-report/
    Belinda Weaver
    @weaverbel
    Raised an issue about using Markdown in our data intro LibraryCarpentry/lc-data-intro#51