Fowl Escapades
2000 Zindi Points
Southern African Bird Call Audio Identification Challenge
145 data scientists enrolled, 20 on the leaderboard
16 January—20 April
Ends in 10 days
EDA in R
published 9 Mar 2020, 10:16

I've decided to take the R requests as a handicap challenge. This is still a work in progress, but meanwhile you are welcome to keep an eye on my git repo.

  1. My intention here is to start off with just EDA in R on a few sample files. Then, you can run that EDA on the test and training files if you wish.
  2. After that, I might do a full starter code in R segment as I have most of the information around this gathered already.
  3. When my code is finalised, I will publish a blog post with more information relating to audio signal processing and R.

This is a pre-emptive strike in case anyone wants to start off in R already. I will supply some blog posts once I have finalised my code, but what you can already see in the code segment is how to display the waveform, and how to obtain the spectrogram.

Hope this helps!

My git repo for Fowl Escapades is here. This is the R code:

https://github.com/DevilEars/Zindi/tree/master/Fowl%20Escapades/R

edited 1 minute later

OK dweebs, I did a blog post with some background domain knowledge if you are new to audio analysis. I am not an engineer, I just fix their code for a lving *cough.

I'll see how this week goes, I would very much like to see my PC catching fire due to me trying this sort of thing in R instead of something decent so what I'm saying is I may finish an R starter code sample. But don't hold your breath, feel free to proceed.

https://dev.to/devilears/zindi-fowl-escapades-challenge-eda-of-audio-data-in-r-ejj