I wanted to introduce a team doing some research and product design for the Jupyter.org website, as well as give a space to discuss the needs related to the website and see if there are any community members willing to speak with the the team and help them understand the ecosystem a bit better. The website has some design and tech elements that are depreciated, I hope we can use this thread as a forum to discuss what the website ‘ought’ to be and ‘ought’ to accomplish.
There’s an issue open on the jupyter.github.io repo: https://github.com/jupyter/jupyter.github.io/issues/331
Over the next couple weeks the team will probably start with some contextual interviews, to understand the ecosystem in depth.
I would be interested in collaborating.
Also, the following post has some good ideas and suggestions in this area.
One quick question: I’m a little bit confused by your mention of JupyterLab in the title. Are you talking about the JupyterLab documentation, or the website at jupyter.org (which is for the whole Jupyter community, of which JupyterLab is one piece)?
That’s my mistake, I made the changes right after the JuptyerLab weekly meeting, it was on the brain. Fixed now.
Thanks for the link, @lresende, we’re keeping our eyes on that thread for reference as we begin the process of figuring out the website’s IA. I am the lead researcher on this team, and as Tim mentioned above, we are conducting contextual interviews for the next 2 weeks. We’re very excited to hear from contributors like yourself, so if you don’t mind giving us an hour of your time, we would love to set up a call with you!
I’m happy to chat with y’all if you like.
Thanks @choldgraf ! I’ll shoot you an email. If anybody else is interested (willing? ) please feel free to email me firstname.lastname@example.org to get the schedule set up. Or just comment here and we can coordinate.
Maybe you can post the questions you have here and let people reply in discourse or by email? I think that more people would have time for that then finding an hour and a timeslot etc.
For right now, we want to start with one-one-one conversations that can allow us to dive deeper into certain topics and have a more free-flowing discussion. We are less interested in specific Q&As and more interested in listening and learning from anything you have to say about Jupyter. Later on, we will definitely be asking for more specific feedback through discourse as we go. We understand that everyone is spread far and wide and extremely busy, but if anyone is able to carve out 45 minutes for an interview, we would really appreciate it!
Hey everyone! We’re looking for more members of the community to do interviews with. If we’re quick, it can be done in 30 minutes. We would love your help on this.
I’d be happy to do an interview.
Also, I tweeted this out here: https://twitter.com/zrsailer/status/1118939825917181953
Hey everyone! We’re making great progressive doing interviews and understanding what jupyter.org means to every member of this community. If you don’t have time for an interview, we’d really appreciate it if you completed a brief survey for us. It’ll take 5-10 minutes of your time. Jupyter Survey
I think it would be great if the new jupyter.org would be organised so that people who arrive there with a thought like “I need to solve problem X, I wonder if Jupyter has something for that?” can find an answer to their question.
Right now the landing page presents solutions (the notebook and the hub) not a problem the reader has/wants to solve.
For example a task a visitor might want to complete is “I want to provide a central place where a small team can login and run their notebooks.” which then points to The Littlest JupyterHub.
“I want to write a text book like webpage to accompany my lecture course, lots of examples are written in notebooks already, how do I create a webpage from them?” which points to Jupyter Book.
“I just receive a notebook by email, how do I view this?” which points to nbviewer (??).
“Someone emailed me a .ipynb file, what do I do with it?”
These are great scenarios @betatim! We’ll look at incorporating some of them into our usability testing. We’re going to evaluate how the website addresses these scenarios.
Definitely! You make a really good point, and we have noticed that, too in our evaluations of the website. Actually, since we’re on the topic, our team is putting together a usability test for potential users. They will be unmoderated but recorded with audio and instructions to narrate their thought process, so it will provide us with great qualitative data on how they approach and navigate the website. Here are a few hypothetical/situational tasks that we were considering. (Note that we are targeting non-users from various backgrounds.) What’re your thoughts on these?
- Imagine you’re a student and looking for a new computational notebook. Find out what Project Jupyter’s core values are and what it’s about.
- Imagine you’re in the Finance department, and you’ve heard about Jupyter to analyze large datasets.
Try Jupyter. Now install Jupyter.
- Imagine you’re a professor who wants to setup Jupyter notebooks for your class. You want them all to use the same version of the software and utilize the same dataset. Find how you would do this.
This would be similar to your suggestion: You want to provide a central place where a small team can login and run their notebooks. How would you do this?
- Imagine you’re a developer, and you want to submit an issue or a bug fix for JupyterLab. Find how you would do this.
- Imagine you’re a corporate sponsor and want to donate to Project Jupyter. Find how you could do this.
And then to add your suggestions here:
- Imagine you just received a notebook by email. How would you view this?
- Imagine you are a professor and want to write a textbook-like webpage to accompany your lecture course. Lots of examples are written in notebooks already. How would you create a webpage from them?
- Imagine someone emailed you a .ipynb file. What can you do with it?
Also, going back to what you said about just seeing the solutions on the landing page, we are actually working on that through our interview analyses. We are wrapping up our interviews this week and wanted to say thank you to everyone who has reached out or helped us in any way so far! From these interviews, we will be extrapolating Jupyter’s greatest strengths in order to represent the products in a more compelling way. And a large part of effective storytelling and drawing engagement is in tying our solutions to the user’s potential problems, so we will definitely also consider your hypothetical scenarios when organizing the information architecture
We are looking to launch our usability test by the end of the week, so everyone, feel free to give us feedback for the next few days!
Awesome! Sounds like you have the ideas and a plan to make it happen!
One thing I recently started thinking/realising is that in the grand scheme of things no one knows what Jupyter is. There are a few million notebooks on GitHub which is a huge number, but compared to the number of Word documents or Excel spreadsheets it is probably just “a rounding error”. You can ask random people on the street and they probably know what a Word document is, less will know what an Excel spreadsheet is and no one will know what Jupyter is.
The other thing I think is true that even though people know what word and excel are, what they want to do is not “use word to write a document” or “make some pivot tables in excel”. What they want to do is write a flyer for their lemonade stand or create that three page report for their boss or compute the sales numbers.
Based on all that: I started thinking we should formulate the tasks without using the word Jupyter or notebook or similar. It is super tricky to do but worth it (I guess) .
The Jupyter is community is thriving and growing and so is the number of components/sub-projects around Jupyter. IMHO, Jupyter.org will not scale if it tries to answer all the questions or describe all use cases of all it’s related components/sub-projects. We should also look around and try to learn from other umbrellas organizations and how they provide overall information about the community and redirect users to the sub-projects website which then provides guidance/responses to these questions in the scope of that sub-project.
I’ve got another suggestion for a task: imagine you’re a (technical, semi-technical or non-technical) manager, you’ve never heard of Jupyter before but it was mentioned by someone and you want to have some idea of what they’re talking about before discussing it further and allocating resources to it.
For that, an elevator pitch slide deck might help (mostly to be used by people trying to “sell” Jupyter to their company / deciders), as do showcases (look what type of problem Jupyter can solve, with the following benefits and positive outcomes).
These use cases are super helpful. @cmbui do you think you could make a Google Doc that we could copy these use cases to, and allow others to dump their ideas. We can use these both for user testing, and a way to validate future designs.