WideNoise Dumps for Research Purposes


For research purposes we offer a dataset of the WideNoise database in form of SQL dumps to interested people. Before you get access to the datasets, you have to sign up our license agreement and send it via fax to our office. Additionally, we would like to ask you to subscribe to the WideNoise-Research mailing list. Upon receipt of your faxed license agreement, we will approve the subscription request and in the welcome mail you will get instructions on how to access the dataset.

On this page you can download the dumps as compressed tar archive archive. A README describing the format of the files is contained in each archive. Please note that the easiest way to work with the dumps is by using a MySQL database. Detailed information on the table structure can be found below on this page.

We are quite interested in results you got with the help of this dataset. Therefore, please inform us about your publications. Concerning citing this data in publications, refer to the following reference:

Martin Becker, Saverio Caminiti, Donato Fiorella, Louise Francis, Pietro Gravino, Mordechai (Muki) Haklay, Andreas Hotho, Vittorio Loreto, Juergen Mueller, Ferdinando Ricchiuti, Vito Servedio, Alina Sîrbu, and Francesca Tria: Awareness and Learning in Participatory Noise Sensing. PLOS ONE, 8(12):81638, 2013. DOI: 10.1371/journal.pone.0081638. Download: PDF.

If you want to refer to the system, please use the following publication:

Martin Becker, Juergen Mueller, Andreas Hotho, and Gerd Stumme: A Generic Platform for Ubiquitous and Subjective Data. 1st International Workshop on Pervasive Urban Crowdsensing Architecture and Applications, PUCAA 2013, Zurich, Switzerland - September 9, 2013. Proceedings, pp. 1175-1182, ACM, New York, NY, USA, 2013. DOI: 10.1145/2494091.2499776. Download: PDF.


File Size Description
2015-01-29.data_widenoise.sql.gz 2.40 MB WideNoise dataset from January 29th, 2015.
WideNoise_License_Agreement.pdf 526 KB WideNoise license agreement (version 2015-01-30)

Dataset Description

The datasets have been created using the mysqldump command of a MySQL database. The CREATE statements for the corresponding tables can be found in the corresponding dump files together with the INSERT statements which insert the data into the database.

We offer two dataset dumps contained in the following files that you can download above after signing the corresponding license agreement.

The WideNoise table contains a shared set of common columns that store shared information across our system as described bellow.

Name Type Description
id bigint(20) unsigned The data points identifier
meta_user_id varchar(64) The hashed unique user identifier. Cells contain NULL in case the user has no user account at our web application .
meta_device_id varchar(64) The hashed unique device identifier.
meta_timestamp_recorded datetime Date and time when the data point was recorded on the smartphone.
meta_timestamp_received datetime Date and time when the data point was received at our server.
geo_lon double The longitude coordinate of the data point's location.
geo_lat double The latitude coordinate of the data point's location.

The WideNoise dump has the following additional data columns that contain the actual measurement:

Name Type Description
id bigint(20) unsigned The data points identifier
data_average double The average dB(A) value shown to the user for the recording.
data_user_estimate double The guessed db(A) entered by the user during recording. Cells contain NULL in case the user didn't used the guess game feature.
data_duration double The duration of the recording in seconds. The default duration of a recording is 5 seconds and can be extended to 10 and 15 seconds.
data_perception_feeling double How lovely was the sampled noise. Ranges from -1 (love) to +1 (hate).
data_perception_disturbance double How disturbing was the sampled noise. Ranges from -1 (calm) to +1 (hectic).
data_perception_isolation double How crowded was the environment during the recording. Ranges from -1 (alone) to +1 (social).
data_perception_artificiality double How artificial was the recorded noise. Ranges from -1 (nature) to +1 (man-made).