mySoundscape: Analysis and Visualization of Audio Samples

In this research project, a mobile device was used as a sensor, to study noise levels in my daily environment. The objective was to look at how much noise pollution I experience, as I go about my day to day life.

For data collection, I developed an App for the Android platform, making use of the audio libraries for taking audio samples on an Android phone. The app ran as a background process and audio samples were stored on the device, until transferred, over usb cable, to a laptop.

Data collection included one-second recordings, scheduled every minute, every day, for several weeks.

Audio analysis, and visualization of the results, was done in Python.

Technologies:

• Android -- for data (audio) collection, with an Android application installed on my phone.

• Python (with Numpy, Matplotlib, pymongo libraries) -- for data analysis and visualization

• MongoDB -- for persisting results, to support multiple sessions of analysis

Observations:

Because the device's microphone was designed to optimize for voice, and to reduce detection of background noise, these results show the most conversational times (which may not be the most noisy times) in my daily environment.