WOLOF ASR data on urban transport

This is document is to present a wolof speech recognition dataset collected and prepared by
BAAMTU Datamation (a senegalease company focused on using data to help companies to
leverage AI and Big Data ).
Motivation
In a country(i.e Senegal), where about 50% of the population is illiterate, using existing
technologies and applications that are designed to be used by people who can read is very
difficult for those people. Our Goal here is to use ASR technique on WOLOF to help the illiterate
persons to interact with apps with just with their voice in a language they can already speak (i.e
WOLOF).
We chose the urban transport use case for two reasons :
● Many urban transport users can’t read nor speak french, so they can’t interact with
existing apps that help passengers to find a Bus for a given destination.
● There is already an existing app in SENEGAL (i.e WeeGo) which help passengers to get
information about urban transport, so the goal here is to build an ASR model that will be
plugged into the App so illiterate people will be able to use it(the resulting model is
actually used by the WeeGo App).
Wolof is the most used language in SENEGAL.