[BlindMath] Ideas for data sonification

[Robin Williams]

Hi all,

Apologies for the cross-posting.

Some of you may be aware of the sonification R package, which can be used to create soundscapes of data. The package was primarily written by a friend and former colleague of mine, Stefan Siegert, with some input from me. Stefan and I are planning to publish an update to the package, hopefully alongside a paper regarding accessibility in the R Journal.

I'm reaching out to ask for any ideas the community may have for things they would like to see added to the package. Here are two items I am hoping to add, but other suggestions would be very welcome.
1. Include a 'audioloess' wrapper function which I use regularly. I use this function to produce Loess fits to x-y scatter data, and then call the sonify package to sonify the output. This is currently my primary use of the package.
2. Add functionality to communicate a measure of noise, or spread, of x-y scatter data. A Loess curve only enables me to understand the shape of the Loess fit, but getting a handle on the spread of points in different regions of the x axis is more difficult. The current idea is to overlay white noise on the soundscape, estimate the size of a confidence interval using quantile regression, and then map the volume of the white noise to the size of the confidence interval.

Please reach out to me either on or off list with any ideas you may have, big or small.

With best wishes,
Robin