AFEC Visualizers

This repository contains visualization experiments for AFEC's classification models using plotly dash.

Explorer

Shows high-level features of an afec database in a simple table. Clicking on an entry shows the sample waveform and plays the audio file.

Usage

python explorer.py PATH_TO/afec.db

Then open the dash server's URL in your browser. This usually is http://127.0.0.1:8050/

Dependencies

Python3 with the following pip modules:

• pysqlite3, plotly, dash, pydub, just_playback

Classification Cluster

Creates a 2d t-SNE cluster from the afec high-level classification data. Clicking on a point in the plot shows the samples detailed classification scores and the sample waveform and also plays the audio file.

Usage

python classification-cluster.py PATH_TO/afec.db

Then open the dash server's URL in your browser. This usually is http://127.0.0.1:8050/

Dependencies

Python3 with the following pip modules:

• pysqlite3, sklearn, densmap-learn, pandas, numpy, plotly, dash, pydub, just_playback