2022-01-28 18:43:12

I'm in a data science class that introduced Jupyter Notebook. After a quick Google search, it looks like it's not accessible enough to be worth using. Are there alternatives anyone can recommend?

This assignment in particular is all about outputting graphs and becoming familiar with plotting where my grade will depend on the output having certain visual features that I cannot check for anyways, but I want to explore this as much as I can before figuring out how to move forward.

What game will hadi.gsf want to play next?

2022-01-28 19:34:03

google colab is accessible to an extent.
you also can use jupytext and turn your python script into jupyter notebook

2022-01-30 17:33:34

Hi, I recently graduated University and had some data science courses where jupyter notebook was used as the de facto IDE. The accessibility story of using jupyter notebook back then was really not great, but I managed to mash together a workflow that allowed me to get an A in that class. And I also had to make graphs, which wasnt great being completely blind, but doable.

Fortunately, the accessibility story for jupyter is way way better now! So you will have a way easier time than I did.

Simply use VS code together with the python extension. Maybe there was another extension that was necessary as well, but in either case when you open a notebook (.ipynb) in VS code it will recommend the correct extension to install. And just like that presto! Accessible jupyter a la cart!

2022-01-30 21:52:53

Thanks for the tips - I'll have to plug away at some things and see what works best.

@3, how did you manage to independently produce graphs and ensure they maintained the features that needed to be present to receive points on the assignment? I fail to see how this can be done without sighted help.

What game will hadi.gsf want to play next?

2022-01-31 13:57:22

@3:
I'm also curious to know how did you managed graphs.

2022-01-31 17:28:56

Well, it helped that I was sighted earlier so I have seen graphs before, so I knew what I was going for. But I didnt do it independently. Of course I had sighted help. I primarily worked on the assignment at home and every now and then I would ask my girlfriend to take a look and ask her questions to get information of the graph that I had so far.

However, the more familiar I got with the api for numpy and matplotlib, the less sighted feedback I needed and eventually could produce decent graphs with sighted help just at the very end to check it is as I expected

2022-01-31 18:14:55

the thing is, you consider each feature in an axis in your graph, and consider your graphs based on them. you can label them with plt.label, limit them into a range with plt.xlim and plt.ylim, etc. but you need to ask someone sighted to an extent to tell you about them. but it is possible to generate them at least.

2022-02-01 01:59:14

Ah, understood. Generating graphs will be fine as long as I have someone sighted to sanity check.

What game will hadi.gsf want to play next?

2022-02-03 23:22:30

for sanity check yes. but for example, you can plot 2 features and check their relations together. one will be moved in x axis, another will be moved in y axis. this is the main thing that you should consider, then in order to make them appealing, someone sighted should tell you. otherwise you will be fine to an extent.