- Data Science - lists.blindcoders.com

Live stream: So how accessible are (Jupyter) notebooks for screen reader users?
by Tony Fast 24 Jun '24

24 Jun '24

cross posting an event from zersiax on the jupyter discourse <https://discourse.jupyter.org/t/stream-so-how-accessible-are-jupyter-notebo…> happening tomorrow 6/25 on twitch <https://twitch.tv/ic_null> and youtube <https://youtube.com/@blindlyCoding> @ 3pm pst. message reposted below Hi all, Accessibility is one of those topics that is often ignored until it’s too late, at times deliberately, and at other times completely by accident. Unfortunately, the consequences of this are often far greater than most people realize; an inaccessible tool/platform can get companies sued, can get people fired (if their current employer adopts these tools), and can make huge portions of a target audience feel discriminated against and excluded. Notebooks, either built on Jupyter Notebooks, Google Colab, VS Code and what have you, are becoming more and more of a staple within various fields, be it AI, Machine Learning, Datascience, programming education and more. So how accessible ARE these tools, at present? Hi! I’m a fully blind (no sight, no monitor, no nothing) developer and accessibility auditor that likes to create this kind of content, and this is exactly the question I aim to answer in the coming (few) stream(s). Tomorrow at 3 PM EST (noon PST, 8 PM BST, 9 PM CEST), I will hop onto twitch <https://twitch.tv/ic_null> and Youtube <https://youtube.com/@blindlyCoding> in order to start answering this question. We’ll look at Jupyter Notebooks first and foremost, bring in resources the community have created over the years to talk about notebooks in general and their accessibility and talk about the notebooks UI from both a web developer as well as an accessibility expert’s point of view. Along the way, topics like " how would you represent data for a screen reader user?" and “if there are accessibility issues, how can a screenreader user best work around them?” will come up as well. There’s a huge lot to cover, so we may have to split this across several streams but for now, I hope i’ll be seeing some of you there tomorrow.

1 0

nonvisual data science live stream 6/20 @ 9:15AM PDT
by Tony Fast 21 Jun '24

21 Jun '24

hey blind coders, tomorrow, thursday 6/20 @ 9:15 AM PDT we'll be streaming about nonvisual data science on youtube <https://youtube.com/live/uF9VpP-mrSE?feature=share>. tonyfast <https://github.com/tonyfast> will be joined by khalid to talk about nonvisual data science and constructing heatmaps from real data. we'll learn about khalid's experiences in nonvisual data science and open source programming. best, tony

1 1

Re: Dash vs. Streamlit for GUI Development for a Data Science Project!
by Khalid K 28 May '24

28 May '24

Marvin , you can: pip install streamlit Or if you're using Anaconda, you can try: conda install -c conda-forge streamlit The same can be done with Dash. Cheers, K From: hunkinmarvin5(a)gmail.com <hunkinmarvin5(a)gmail.com> Sent: Wednesday, May 22, 2024 7:05 AM To: 'Khalid K' <kaak42(a)outlook.com> Subject: RE: [data-science]Dash vs. Streamlit for GUI Development for a Data Science Project! Hi using jaws 2024. Using visual studio code, python 3.12 and github latest desktop version. Have not used Streamlit. And so a tutorial how to use this would be great and maybe a sampleproject. So then I could then see how accessible it is with a screen reader. So where do I download this and is it a pip install like on the terminal or is it a msi or exe file for windows 11. Have windows 11 pro 64 bit on a hp pavilion 15.6 inch laptop. Running jaws 2024, nvda 2024.1. Marvin from Adelaide, Australia.. From: Khalid K <kaak42(a)outlook.com<mailto:kaak42@outlook.com>> Sent: Wednesday, May 22, 2024 3:29 PM To: datascience(a)blindcoders.com<mailto:datascience@blindcoders.com> Subject: [data-science]Dash vs. Streamlit for GUI Development for a Data Science Project! Hi data scientists, I hope this message finds you well. I want to gather some insights on using Dash or Streamlit for GUI development. Specifically, I'm interested in the following: 1. Ease of Use with Screen Readers: Has anyone used Dash or Streamlit with a screen reader? Which one is more accessible and user-friendly? 2. Learning Curve: Which tool requires less time to learn and implement effectively? I have a few sighted friends who speak highly of Streamlit, mentioning its simplicity and ease of deployment -they don't use screen-readers though! My goal is to develop a GUI interface that can: Generate graphs Allow selection from a list of cryptocurrencies Provide scenarios like predicting next month's prices if I sell Load results and graphs quickly upon selection Your experiences with either tool would be greatly appreciated. Cheers, Khalid

2 2

Dash vs. Streamlit for GUI Development for a Data Science Project!
by Khalid K 22 May '24

22 May '24

Hi data scientists, I hope this message finds you well. I want to gather some insights on using Dash or Streamlit for GUI development. Specifically, I'm interested in the following: 1. Ease of Use with Screen Readers: Has anyone used Dash or Streamlit with a screen reader? Which one is more accessible and user-friendly? 2. Learning Curve: Which tool requires less time to learn and implement effectively? I have a few sighted friends who speak highly of Streamlit, mentioning its simplicity and ease of deployment -they don't use screen-readers though! My goal is to develop a GUI interface that can: Generate graphs Allow selection from a list of cryptocurrencies Provide scenarios like predicting next month's prices if I sell Load results and graphs quickly upon selection Your experiences with either tool would be greatly appreciated. Cheers, Khalid

1 0

Notebooks for all May meeting 5/16 @ 9am PST
by Tony Fast 22 May '24

22 May '24

hey blind coders, join us this thursday at 9am pst to discuss accessible coding experiences at the notebooks for all community meeting. our plan for this meeting is to hear about your access experiences and tricks to participating in data science and programming. we want to hear the good, the bad, and the ugly. you can read a full agenda on the notebooks for all github discussions <https://github.com/Iota-School/notebooks-for-all/discussions/109>. please respond there or this mailing if you have any questions or special access needs. feel free to invite your friends, allies, and accomplices. *apologies for the last minute notice.* a special thanks to the iota school <https://iota-school.github.io/> for providing resources to host this event. best, tonyfast <https://github.com/tonyfast> -- Tony Fast Github <http://github.com/tonyfast> | LinkedIn <https://www.linkedin.com/in/tonyfast/> | Homepage <http://tonyfast.github.io>

2 3

OCR Tools
by Nikhil Vohra 20 Apr '24

20 Apr '24

Dear all, I am trying to find a simple OCR tool to run locally. I came across Tesseract <https://tesseract-ocr.github.io/>, and I was wondering if anyone has experience using it, thoughts about it, or suggestions for OCR in general. Note: NVDA has built-in OCR (triggered with NVDA+R), but this only works for the current navigator object; I am trying to find a tool that works for entire documents. Thanks, Nikhil

2 4

Inquiry on Weka Accessibility for VI Data Scientists
by Khalid K 15 Apr '24

15 Apr '24

Hi Data Scientists, I'd like to inquire about the accessibility of Weka (https://ml.cms.waikato.ac.nz/weka/) when using a screen reader. Does anyone have experience with its usability? Additionally, I'm curious to know whether it's essential to use Weka, or if similar productivity could be achieved using Python and libraries like Pandas, Seaborn, and scikit-learn. I look forward to any ideas or suggestions you have for enhancing productivity for data science with screen-readers. Kind regards, Khalid

3 3

Sonification.
by Gabriele Battaglia 08 Apr '24

08 Apr '24

Hi guys. My first question has to do with a subject I care a lot, sonification. I tried astronify but I'm not very happy with it. In my opinion it lacks many things like stereo capability, using different waveforms, sonify more than 1 serie of data at the same time and a usless complex grammar. I would like to try something more complete and maybe easier to use. Do you know something like that, to import as a python library? Thanks. Gabe. -- Gabriele Battaglia (IZ4APU) --... ...-- -.. . .. --.. ....- .- .--. ..- - ..- . . Sent from my Giant Desktop PC

3 2

Reading formulas with NVDA.
by Gabriele Battaglia 06 Apr '24

06 Apr '24

Hi, What's the best addon to read math in web pages or document? I've found MathCAT and Access8Math, in the NVDA addon store. Which one is most complete or still mantained or better somehow. Thanks. -- Gabriele Battaglia (IZ4APU) --... ...-- -.. . .. --.. ....- .- .--. ..- - ..- . . Sent from my Giant Desktop PC

1 0

Enhancing Sonifications with Spatial Effect
by Nikhil Vohra 02 Apr '24

02 Apr '24

Dear all, I recently wrote a script (attached at the end of this email) that, for each .wav file in the folder/directory where the script is run: 1. smoothly pans the audio from left to right over its full duration for spatial effect (such that x is mapped both to time and space). 2. exports the file as a .ogg file, saving memory (without noticeable loss in quality). In order to make use of the script, you will have to save sonifications as .wav files after generating them. Run `my_sonification.write(file_name.wav)` after running `my_sonification.sonify()` to save the sonification as a .wav file that can be played at a later time, shared with others, etc. Replace 'my_sonification' with the relevant variable and 'file_name' with whatever name you choose. Note that Web browsers and audio players should be able to handle the .wav file on your system. The script also has two additional dependencies before it can be used: 1. FFmpeg (Use `conda install ffmpeg`) 2. PyDub (Use `pip install pydub`) Once you've installed the dependencies, navigate to the folder that contains your .wav files (using `cd path_to_folder`), start your IPython session by typing `ipython`, and run the script using `%run pan_and_compress.py`. If the script runs successfully, then all of the .wav files in the folder should have been converted to .ogg files that have spatial audio (panned across the left and right channels) and that are much smaller in size than the original .wav files. You can then open the .ogg files in a Web browser or audio player to hear the audio. Using headphones can make the sonification sound more immersive as the sound moves from left to right as it plays. Based on the perceived location of the sound at any given point, you can estimate your position in the sonification (e.g., you hear a peak halfway through the audio (in the center) or about a third of the way into it (to the left)). If you would like to play your new .ogg file in the IPython command line directly (to avoid the need for other programs), use the following steps: 1. Run `import subprocess` 2. Run `subprocess.run(['ffplay', '-autoexit', 'file_name.ogg'])` Note that you can hit the Escape key on the top left of your keyboard to stop playback at any time while the sonification is playing from your terminal using the above steps. Consider using headphones however/wherever you choose to play the file. You will find 1) the script 'pan_and_compress.py' and 2) 'sample_spatial_sonification.ogg' attached to this email. You can listen to the sample spatial sonification to get an idea of the left-right panning. Remember to move the script into the folder where you would like to use it. Please let me know if you try it out and/or if you have any questions. Best, Nikhil P.S.: The sample sonification is of the mathematical function the cosine of *x* squared, *y* = cos(*x***2), from *x* = -6 to *x* = 6. Python uses two stars to indicate exponentiation.

1 0