PyDoA Python Package¶
The PyDoA package provides functionality of processing and visualizing DoA data. It provides the following main features:
- The ability to map degrees to speakers
- The ability to extract speaking time for each speaker
- The ability to generate window-wise speaking time
- The ability to generate interaction network
- The ability to generate a web-based dashboard for visualizing the DoA data in an interactive manner.
Quick Start¶
A sample data field is provided in the samples directory. Following examples are using that file.
Plotting smoother degree distribution¶
>>> from DoaProcessor import DoaProcessor
>>> d = DoaProcessor('./samples/sample.csv',4)
>>> # You can tweak value of sigma to change the smoothness
>>> d.getPeakDegree(sigma=2,group='group-1')
Degree distribution plot will help you in understanding the directions for each speaker. For instance, the above figure shows that highly captured directions are around 0, 100, 180,300 degrees.
Extracting speaking time for each speaker¶
In order to extract the speaking time, you need to specify the corresponding degrees for each speaker. This knowledge can be gained from the sitting arrangement of speakers around microphone arrays. You can also learn about it by manually exploring the degree distribution.
>>> from DoaProcessor import DoaProcessor
>>> d = DoaProcessor('./samples/sample.csv',4)
>>> # manually setting directions for each user
>>> d.setDegreeForSpeaker(group='group-',[30,104,194,315])
>>> d.getSpeakingTime(plot=True,time='sec',group='group-1')
Generating social network¶
>>> from DoaProcessor import DoaProcessor
>>> d = DoaProcessor('./samples/sample.csv',4)
>>> # manually setting directions for each user
>>> d.setDegreeForSpeaker(group='group-',[30,104,194,315])
>>> d.drawNetowrk(group='group-1')
