SoundMap: A Visualization tool to explore multi-attribute sound data

Demo: https://youtu.be/8jrC7JqcD4Y

Data

https://drive.google.com/file/d/16U_h_fT6FOWxaMhx-mQXEai6l_KmxnuG/view?usp=sharing

Description

SoundMap is created with ReactJS. We added the library d3 and material-ui (and other supporting libraries like axios for api communication). We use a Header component as the toolbar to show soundmap logo and the buttons for manipulating the dashboard. The Dashboard is the main page where we show all other chart components like Overview, SummaryView and DetailedView. We mainly use d3 to create beeswarm charts and bar charts from scratch in the Overview and SummaryView components. We track click events on the dots for showing the detailed view and mouseover events for showing the tooltip over the dots or bars. The detailed view is rendered only when a dot is clicked. We used a skeleton loader as a loading svg whenever the detailed view is re-rendered based on the dot clicks. We added a filter panel which uses material ui components and icons for the form elements like checkboxes, dropdowns, input fields etc. and uses css for showing/hiding the panel. For the multi-view feature, we use react states and CSS to re-arrange the charts in a juxtaposed manner.

SoundMap uses a React client and a Flask server. The server and client communicate via REST API. We use the library axios in ReactJS for calling the backend endpoints. There are two backend endpoints, /api/generate-spectrogram and /api/generate-waveplot. We use librosa and matplotlib to render the spectrogram and waveplot based on the request body parameters (file_data and sample_rate) of a particular sound file. The spectrogram and waveplots are sent as base64encoded image to the client. The client renders this base64 image using the <img /> tag.

Below is the setup instructions of the app server and client. Use two terminal tabs to run the client and server simultaneously.

Setup

Server

Creating virtual environment

Using conda

• conda create --name venv
• conda activate venv
• conda install --file requirements.txt

Using pip

• Create a virtual env by installing virtualenv package from pip
• virtualenv venv
• source venv/bin/activate
• pip install -r requirements.txt

Running the app

• ./bootstrap.sh

Client

• Make sure to use node version >=12
• Install yarn using npm i -g yarn
• Install packages using yarn install
• Run the app using yarn start and view it in http://localhost:8080/