Hey Patrick,

 

Thank you for sharing this information about Streamlit and its approach to accessibility. It’s helpful to see that they have some considerations in place, but it’s a bit concerning that there hasn’t been a response from the Streamlit team since June 2021 on this topic.

 

I’ll explore different options for the GUI in this project. If I decide to try Streamlit, I’ll make sure to share my experience with the group. A friend of mine, who is good at CSS, mentioned that he will use Flask for the GUI. This might be a faster and easier option to implement while meeting the requirement!

 

Congratulations on the new job!

 

Kind regards,

Khalid

 

From: Patrick Smyth <patrick@iotaschool.com>
Sent: Saturday, May 25, 2024 4:21 PM
To: datascience@blindcoders.com
Subject: [data-science]Re: Dash vs. Streamlit for GUI Development for a Data Science Project!

 

Hey Khalid,

 

I'd try looking at Streamlet, I found this comment that suggests they do actually think about accessibility:

 

```

Thanks for the question. We haven’t done an official accessibility assessment. That said, we write our frontend code with accessibility in mind, using a widget component library 31 that is accessible, looking at contrast ratios when designing, etc.

 

But we’re aware this is by all means not enough! That’s why we have a roadmap item to officially assess accessibility and address issues that arise. It’s important to note we don’t control user themes, plotting libraries, and custom components, and a lot of the accessibility of actual apps will fall on users who are writing these apps.

---

 

That's from here: https://discuss.streamlit.io/t/508-compliance-for-streamlit/13579/2

 

Khalid, I owe you an email, apologies. A new job and baby mean I'm a little underwater.

 

 

Patrick

 

--

Patrick Smyth, PhD

Chief Learner

Iota School

GitHub | Twitter | LinkedIn

 

 

On Wed, May 22, 2024, at 2:16 AM, Khalid K wrote:

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@gmail.com <hunkinmarvin5@gmail.com>
Sent: Wednesday, May 22, 2024 7:05 AM
To: 'Khalid K' <kaak42@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@outlook.com>
Sent: Wednesday, May 22, 2024 3:29 PM
To: 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

_______________________________________________

Data Science mailing list -- datascience@blindcoders.com

To unsubscribe send an email to datascience-leave@blindcoders.com