Link to http://nipy.org/workshops/

[arokem]

Where should we provide a link to the workshops?

[vsoch]

How about an events tab?

[arokem]

I am not sure that we want to keep this up-to-date with events as they happen and expire.

I was thinking of linking to that page as a clearinghouse for workshop materials (which is what it looks to be designed to do)

[satra]

nipype is the only one who has used it thus far, but i used it for:

• announcement
• workshop itself
• post workshop clearinghouse

we almost need a conda forge for teaching materials and notebooks, as they keep evolving - perhaps time to think about one.

[satra]

edx-forge anyone?

[vsoch]

🍿

haha just kidding, I really like this idea! I haven't used the real conda forge a ton - @satra could you explain what edx-forge would do/be? Either there is something out there that we could use, or we could make a new one :)

[vsoch]

I think things like workshops (at least in the current format) would best be put under the blogroll (tumblr if I remember correctly). There are so few posts on there, any new posted events and what not would stay at the top, and subscribers would be notified.

[satra]

this is going to distract from @arokem's original question, but i'm too lazy to start a different thread right this minute. if it gains traction amongst us, we can bring a different thread and bigger community to bear.

education is just like software imo.

• each module has dependencies/requirements effectively creating a directed acyclic graph (much like a workflow
• each module needs tests
• each module evolves and dependencies evolve over time
• each module has different scales (small teaching elements, big teaching elements)
• each module determines what functionality it provides and what the complexity of the api is.

now if we only had a way to present materials in multiple mixed formats in a package (slides, videos, notebooks) and figured out a way to run a test (container, form, notebook, etc.,.) and evaluate it (we can create a course).

just like conda forge, courses would have:

• maintainers, and they can bring different perspectives angles
• evolve with versions

[vsoch]

clarification question - the dependencies/requirements and tests refer to things that are relevant to the student? eg: the student must have dependencies/requirements by way of completing other courses, and finishing to get those credits they must pass tests?

The only issue I see is that the dependencies shouldn't necessarily be taking other courses - eg, someone who is good / self taught at something, but hasn't taken a course.

I think the first step would be figuring out the data structures /formats (and likely Github would be involved) and then figuring out how to hook in the interactive bits to that. Have you seen Github Classrooms? https://classroom.github.com/#getting-started-video It is a good starting thing to think about, because it lets you create classrooms that are basically clones of another teaching "template" - meaning a version of a class is tied to the instance when it was taught. And there is nice integration of being able to collect assignments and interact with the students.

[vsoch]

I think the hardest thing would be organization, and labeling things, and then placing content. Maybe we could just start with wikipedia --> http://datamining.typepad.com/data_mining/2007/10/visualizing-wik.html and a forge would be some subset of pages and topics from there, with additional media from the instructor like slides, python notebooks? I mean, doesn't everyone end up there anyway? lol

[satra]

@vsoch - the github classroom bit is primarily for code based assignments. what we envision are topics that may not have code assignments (even though i'd like to start in the space of data science, where data and code would be required). we can definitely use it for that purpose. but the meta layer for students and grading would be separate from the layer for content and evaluation questions.

for several things we have used the software carpentry style template to create content repos (e.g., https://github.com/neurohackweek/, https://github.com/repronim/), but they are very static, doesn't quite have a system for evaluation and puts a fair bit of effort on the part of the student.

i quite like the conda forge style of feedstocks. it really distributes the problem of dependencies to constructors of the feedstocks. the part where it diverges and where we need to have some thought is that the tarball underneath a conda forge feedstock has a consistent hash and while the numpy feedstock has versions and builds, there are no two approaches numpy. education can have many perspectives, and i'd like to think about a way such perspectives can be brought to bear in this scheme.

the reason wikipedia doesn't work is because it's static. part of the reason ariel asked this question above where to point to the workshop was because of the various materials that came out of the workshop. without a software overhaul, wikipedia wouldn't be able to maintain such content effectively. wikipedia is great for encyclopedic content, but not good for pedagogy. it's a look up table rather than an interaction.

[vsoch]

I was thinking about wikipedia not as a solution, but as a starting skeleton for topics. Eg, a classroom would select some subset of topics (wikipedia pages) and some level of branching off of that structure would define the "topic tree" and the instructor wouldn't necessarily need to use wikipedia, but could add layers of content on top of it, associated with some chunk of the tree. It's a cool approach because then, given the integration is active in the browser, someone could go to any wikipedia page and see the associated courses, and people with two "different" courses would see the overlap.

I'm going for a run - will give some thought about this.

[vsoch]

Here's an interesting visual (albeit static in this example) way to put together resources for many unlike things! And good links / resources too: https://www.datacamp.com/community/blog/data-science-periodic-table#gs.4oS99a0

[arokem]

Very interesting! Reminds me a bit of https://nodeschool.io/ For the time being, though, I think that it would be good to at least link to the many different workshops that teach neuroimaging in Python, and their teaching materials. I think that a "workshops" tab in the main menu might be generic enough for us to link to that page. #40 BTW - where is the source for the page at "nipy.org/workshops"?

…

On Wed, Apr 12, 2017 at 8:03 PM, Vanessa Sochat ***@***.***> wrote: Here's an interesting visual (albeit static in this example) way to put together resources for many unlike things! And good links / resources too: https://www.datacamp.com/community/blog/data-science- periodic-table#gs.4oS99a0 — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#39 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAHPNqJgcpZGFxzVCF15dA__j_tFlhTYks5rvZBzgaJpZM4M6cm0> .

[vsoch]

I don't know, but I think it looks like Ubuntu, circle 2011 :)

[satra]

@arokem - https://github.com/nipy/workshops/tree/gh-pages/

it's the index.html file there.