Elevator Annotator

Elevator Annotator is an Information Science master thesis project which takes place in both Vrije University Amsterdam and Netherlands Institute for Sound and Vision.

About the Project

The idea is to find out and analyze if physical crowdsourcing works for annotating a set of audio files. The dataset is compiled from a set of public domain songs from Europeana’s collection.

What makes physical crowdsourcing different from the normal crodwourcing is that instead of waiting users to do the task, physical crowdsourcing approaches the users and ask them to do the tasks. The concerns then goes on whether the users are willing to do the task, correctiveness and validity of the task will then be analyzed further along with the results.

Elevator Annotator combines the approach with pervasive computing source as a device/platform is built to do the experimental task on a Raspberry Pi combined with a set of sensor and audio I/Os.

Getting Started

These instructions will work as a guide to help building a copy of the project up and running for development and testing purpose. The project may be re-used and further developed for other purpose.


Please ensure both hardware and software prerequisites are met with the corresponding versions.

Hardware Requirements

Here is the list needed for building prototype, please note that there are 2 version of prototypes for experiment purpose. The difference is in the input, see note on the list below.

And here is a list of additional features and tools for experimental purposes.

Software Requirements

Here is a list of software and library needed for development purpose along with each versions (version used to run the experiment, please note that older or newer versions may or may not work).


Here is the list of public domain audio files taken from Europeana’s collection, provided with each individual links to the collection.

Song Titles Alias/Filename
Oktāvu etīde song1.mp3
Tautas polka song2.mp3
Dienā jaukā - noskūpstīju Annu laukā song3.mp3
Florentine song4.mp3
Mana dzimtene song5.mp3
Meitenes sirsniņa song6.mp3
Kādēļ tik ilgi vilcinies tu? song7.mp3
Dziedu tev song8.mp3
Serenade iz operas song9.mp3
Serenade song10.mp3

Installation and Development

This section will provide explanation on installing and developing the project for both the hardware and software.

Hardware Configuration

Here is a list of steps needed to assembly the prototype.

Software Configuration

Here is a list of steps needed for preparing and developing the software parts.

Running the Experiments

This section will explain on how to deploy the software and prototype into actions and how the flow for both versions to work in rightful manners.



Here is the workflow for the audio as inputs version:

And here is the workflow for the button as inputs version:

Here is the structure and example of the registered annotation:


A                       B       C       D       E       F       G       H       I       J       K
20170517-11:53:47	1	YES	T	sound9	0.25	trumpet	2	YES	FT	38.0137310028

A video for demo purposes of how the setup looked like for the experiment can be seen on Youtube through the embedded link below: Elevator Annotator Demo


This project is a written and developed by Anggarda Prameswari as a master student in Information Science of Vrije University Amsterdam with the help and guidance of both supervisors:



This project is under the GNU General Public License v3.0 - see the LICENSE.md file for more details.