Data Sonification:
Weather Disasters Declared in the USA: 1970 -- 2021
Fires, Severe Storms, Floods, & Hurricanes
Data from the Federal Emergency Management Agency (FEMA)
What is the frequency trend of the most prevalent natural disasters declared in the USA, since 1970?
For my second data sonification project, I wanted to focus on something that was more sonically intuitive than my last Lord of the Rings sonification exploration. (As fun as it was, I realized it might be difficult for a listener to understand the meaning of the sounds without an informative visual reference or sonic lexicon!) I wanted to create a time-based sonification, and associate numeric value with pitch. A friend of mine suggested the idea of creating a natural disaster-focused project, and after some investigation I found the Open FEMA Data Sets resource.
Considering the most prevalent disaster types in the disaster declaration data over time (fires, severe storms, hurricanes, and floods), I was curious to know what the disaster count trend was over the last few decades. I decided to associate the numeric count of total disasters per year with pitch (higher pitch = higher frequency), and associate the individual yearly disasters with amplitude similarly. The total count is represented by a marimba, and the individual disaster types are represented by weather sound sample tracks. More ambient than anything, you can hear a crackling fire, a rumbling storm, whistling wind, and rushing water to represent fires, severe storms, hurricanes, and floods, respectively.
Below is the sonification, accompanied by an animated visualization I created to pair with it:
Method Details
The R script, datasets, Sonic Pi scripts and output, as well as the sonification audio bounce can be found in the GitHub repository linked at the top of this page.
1. I pulled in the data from the Open FEMA API, then cleaned and analyzed it.
2. I converted the total disaster instance count values into a normalized scale of frequencies that would be audible in Sonic Pi.
I converted the individual disaster count values into a normalized scale of decibel values that would be audible in the mixing process.
3. I used Sonic Pi to record the frequency array that I had assessed in R.
4. In Logic Pro X, I converted the frequency recording from Sonic Pi into a MIDI track so that I could apply different software instruments.
5. I imported sound samples from freesound.org, mixed them, normalized their volumes to have a comparable baseline, and applied the decibel change levels across each track.
6. I exported the resulting audio from Logic Pro X.
7. Using a plot I had generated in R as reference, I created an animated visualization to accompany the sonification, using Procreate and Adobe Photoshop.
1. I pulled in the data from the Open FEMA API, then cleaned and analyzed it.
2. I converted the total disaster instance count values into a normalized scale of frequencies that would be audible in Sonic Pi.
I converted the individual disaster count values into a normalized scale of decibel values that would be audible in the mixing process.
3. I used Sonic Pi to record the frequency array that I had assessed in R.
4. In Logic Pro X, I converted the frequency recording from Sonic Pi into a MIDI track so that I could apply different software instruments.
5. I imported sound samples from freesound.org, mixed them, normalized their volumes to have a comparable baseline, and applied the decibel change levels across each track.
6. I exported the resulting audio from Logic Pro X.
7. Using a plot I had generated in R as reference, I created an animated visualization to accompany the sonification, using Procreate and Adobe Photoshop.
