Les NTICs vocales pour Radio Sikidolo, au Mali

Adama Tessougué, journaliste de Radio Sikidolo travaille à Konobougou, un petit village à 150 km de Bamako, dans une région rurale du Mali. Chaque jour la voix de Adama informe et divertit au moins 50.000 auditeurs dans 39 communes autour de Konobougou. «Informer sans déformer» est écrit sur les murs du couloir de la radio, où nous rencontrons Adama et ses collègues.
AdamaTessougue
«Oubliez le Web ou le mobile, la radio est le plus grand média en Afrique», Mawuna Remarque KOUTONIN, écrit dans son blog Silicon Africa. «La raison est simple, un grand nombre d’Africains sont encore non-alphabetisées…. »

Si on veut améliorer la communication avec le Web, ici au Mali la radio et le mobile doivent être les principales interfaces au Web. Ainsi, l’accès au Web traditionel, basé sur le mot écrit, n’est pas la solution, dans ces conditions rurales Africaines. Ici on aurait plutôt besoin d’un Web vocal.

À Sikidolo il n’y a pas de connexion internet fixe. Adama Tessougué peut vérifier son courrier électronique en utilisant une connexion Internet mobile, mais cela est très coûteux, – il paie une taxe par minute en ligne – pour ça il n’accède le Web que rarement. Communication ici est basé sur la parole.

Nous visitons Radio Sikidolo pour parler avec Adama sur un nouveau service vocal de micro-blogging, qui a été récemment installé pour les radios au Mali. Ce service est nommé Foroba Blon. La voix de l’interface de ce service mobile est en langue bambara ou bomu, permettant aux utilisateurs de communiquer avec le système dans leur propre langue locale.
Adama travaille avec une cinquantaine de journalistes citoyens indépendants, qui distribuent les nouvelles des villages environnants. Ils recueillent des annonces et signalent des nouvelles : les mariages, enterrements, fêtes, accidents, parfois des vaches et des chèvres perdues. Ces blogs vocaux sont envoyés à la radio par mobile.

Le système Foroba Blon a été développé par une équipe de la Fondation Web, la Université VU Amsterdam, North-West University de l’ Afrique du Sud, et l’ONG malienne Sahel Eco, en collaboration avec plusieurs stations de radio au Mali, y compris ORTM Ségou et Radio Moutian à Tominian.

Plusieurs stations radio au Mali sont intéressées à accéder aux nouvelles qui sont diffusées par radio Sikidolo. Ainsi, en élargissant le service de micro-blogging pour les autres stations de radio, et en leur permettant de partager des ressources basés sur la voix, cela pourrait éventuellement devenir ce que nous pouvons nommer le Web de radios africaines.

Un système d’information du marché, à base de voix, pour Zakary Diarra, agriculteur et producteur de miel au Mali

Zakary Diarra, agriculteur, apiculteur et producteur de miel, vit dans le petit village Bokuy-Mankoina, dans le sud-est du Mali. Dans ce village, il n’ya pas de connexion Internet. Toute la communication est basé sur la voix. Pour Zakary les principaux canaux d’information sont le téléphone mobile et la radio communautaire. Depuis quelques années, une nouvelle approche, introduite par ONG malienne Sahel Eco, a changé la vie de Zakary. Il bénéficie maintenant d’un service d’information du marché, basée sur la téléphonie mobile et la radio. L’idée de ce service c’est d’aider les agriculteurs à produire plus de cultures d’arbres, tels que le miel, les noix et le beurre de karité.Mali november 2011

«Malgré les crises alimentaires qui ont affecté le Sahel, ces derniers temps, les pénuries alimentaires sont moins fortes dans des zones à forte densité d’arbres. Petits agriculteurs créent des systèmes agricoles plus productifs et plus résilients à la sécheresse, grâce au reverdissement, qu’ils ont récemment introduit dans leurs terres», explique Chris Reij de l’Université VU d’Amsterdam. Selon Mary Allen de Sahel Eco, les ménages pauvres peuvent améliorer leurs revenus en vendant des produits forestiers, comme du bois de chauffage, des fruits, des noix et du miel. En outre, les arbres améliorent la qualité du sol et fournissent du fourrage pour le bétail.

Pour améliorer la chaîne de valeur des produits forestiers non-lignieux, un système d’information du marché, basé sur la voix, a été construit par les développeurs de l’Université Libre d’Amsterdam (VUA), de la Fondation Web, et de l’Université du Nord-Ouest en Afrique du Sud, en collaboration avec les agriculteurs, Sahel Eco et plusieurs stations de radio du village, y compris Radio Moutian et Radio ORTM Ségou, dans un projet nommé VOICES et une initiative nommée W4RA. Ce système d’information du marché a été surnommé RadioMarché.

Depuis le deployement de RadioMarché, Zakary collecte des offrandes des agriculteurs voisins, les agrégeant dans des offres de groupe. Ce service, qui est tout à fait basé sur la voix, permet aux agriculteurs de diffuser ses offrandes de miel et de beurre de karité sur la radio locale.

J’ai demandé l’avis de Zakary sur cette initiative. Zakary m’a répondu: «Mon revenu de la vente de miel a presque doublé entre 2010 et 2011. Maintenant j’ai plus de stabilité financière, qu’avant ce projet. Je suis en mesure de payer la scolarité de mes quatre enfants et j’ai même pu acheter une charrette et un âne, l’année dernière.
honeybottles

Si je n’avais pas participé dans ce projet, je continuerait comme agriculteur régulier, et j’aurais raté cette opportunité de devenir un entrepreneur en miel. Je suis déterminé à augmenter le nombre de ruches, et je conseille à d’autres dans le village de faire le même, pour que nous en groupe puissions augmenter notre production de miel et le volume de nos ventes, pour répondre à la demande des clients. «Je vois les avantages de ce service RadioMarché, qui a amélioré la confiance et la collaboration entre les producteurs et les vendeurs de miel. Je suis également content d’être une personne de contact pour le système d’information du marché RadioMarché, puisque ce rôle est très apprécié dans mon village. Il y a même des gens qui m’appellent Sozakary , ce que signifie Zakary du miel».

Decentralised Open Data for World Citizens

Brussels – 19 June, Christophe Guéret from VU University Amsterdam presents “Decentralised Open Data for World Citizens” at the Open Data Workshop for policy modeling, citizen empowerment, data journalism. The seminar is organized by W3C in the framework of the Crossover project. The venue is at the European Commission’s Albert Borschette Conference Center, in Brussels. The EC and many national governments are currently publishing open data and supporting initiatives in this direction. Especially the combination with social media data seems to become a hot topic.

Despite the efforts in developed countries, again the developing countries are lagging behind, now in linked open data initiatives. As we watch the geospatial distributions of open data over the world, we see initiatives in Europe and North-America, and a few in South-America. The African map is very empty, with only one small dot in Tanzania. Will this be a second wave of widening the digital divide? Expectedly the interlinking of so many new sources of data wil boost the knowledge economies in the rich and developed world. How will developing countries ever catch up the information society, if they don’t have the infrastructure to access and process all these data? Besides, open data are still very textual, and mainly written in english or a few other world languages. What about the groups that don’t have a tradition of written text? Again these people will stay at the underprivileged side of the digital divide. Is it possible to start linking data from the very beginning, as soon as they are being produced, as is happening now via innovative voice-based systems and newly developed ASR and TTS systems for very small and under resourced local languages…Are there ideas, how we can decentralise ?

The presentation by Christophe Guéret and others, from VU Amsterdam is different from others in this workshop. It is not only socially engaged, but very practical. It talks about real people having real problems. It refers to linking non-existing data, voice-based data that is currently being collected from rural communities in Mali and neighboring countries. Is it too early to talk about linking data when people don’t even have an internet connection? On the contrary, in this paper the authors argue that linked data principles can be applied from the very beginning. The Semantic XO, is developed for “one laptop per child” (OLPC) project, so that children can share and produce data while playing educative games and start creating content in a creative social space. The linked market data from farmers in Mali is an example of local trading activity, that is relevant in the local context, but can be linked with new initiatives in other regions, it is voice-based, and adapted for mobile phones, and it has the potential to connect the farmers to other trading spaces and potential customers…

This presentation is different from other topics discussed here, such as legal issues, huge datasets, governments informing citizens. The initiative presented here is the first one that tackles the problem from the grassroots in developing countries, from the farmers and the school children in remote rural areas. It is not too early to start here…

Downscale 2012 – First International Workshop on Downscaling the Semantic Web

First International Workshop on Downscaling the Semantic Web – Crete, May 28th, in the pre-conference of ESWC 2012. Organized by Christophe Guéret and Stefan Schlobach from VUA and Florent Pigout, from OLPC.

Organizers and participants of the Downscale workshop 2012

The workshop brought together a group of computer scientists, experts in Semantic Web technologies. Although the general trend is upscaling, here, the notion is of downscaling, decentralizing, even reducing the Web of Data, for a special reason. The need for data sharing in the developing world is an important topic, especially the case of rural areas in Africa, and of schools in developing countries. Another area is disaster management, also presented here. In disaster, the amount of data is especially the issue, when rescue teams have to take fast decisions in situations that concern life and death. Here, downscaling is especially necessary to reduce the amount of data, and keep only what is relevant…
For developing countries, three topics were presented, where Linked Data can be applied: (i) Linked Market Data from rural communities in the Sahel, (ii) SemanticXO, linked data to share accross children’s computers that don’t have access to the internet, but still want to share data, and (iii) the Web of Radios, currently only a dream, but one that can become reality, positioning community radio an important future content provider in Africa.

At the conference, key-note speaker Abraham Bernstein argued that we should throw the Semantic Web into the garbage can. The audience was surprised with the answer, that they obviously had expected to be negative. On the contrary, Bernstein explained that he sees the garbage can as the place where theory and practice meet. With the metaphore in mind I thought how it would be to make this highly intellectual Semantic Web community, absorbed in theory, meet rural reality of a farmer community in Tominian, where most of the population cannot read and write. Would this meeting be useful for both groups? I mean, is it possible to bridge two worlds, making both groups learn from each other? Would it yield totally new insights? It was probably not these two extremes the keynote speaker and the conference participants had in mind, but it is an interesting experiment…

Community Radio in Tominian and Segou, Mali.

Foroba Blon is the name of the newest project by the W4RA team from VU University Amsterdam, Web Foundation and Sahel Eco. This is a project on development of voice technologies for innovation in support of free press in developing countries.

The name chosen for this new project, which will support and benefit community radio in Mali is very special, in two words: FOROBA BLON, in Bambara, a Malian language.

FOROBA literally means “big field” or “collective field” but  the signification is “for everyone” BLON = the vestibule where the chief holds his audience. Together they signify the large room or space where everyone has the right to speak in front of the chief; the truth can be  told; but only if you do it respectfully, without insulting anyone.

If you are wondering how to say it its three short equal syllabi. Fo-ro-ba then the accent on Blon (di-di-di-daaaaa).
We think it conveys very well the idea of citizen journalism.

Many thanks to Diarra Fousseyni, from ORTM Ségou, who suggested this meaningful name for our project.

The Foroba Blon project has been selected by the International Press Institute (IPI), with two other projects, out of 376 submissions! To organize this contest, IPI received a grant from Google.

FOROBA BLON started in January 2012 and is funded by the International Press Institute.
.

Foroba Blon, voice-based technologies to support citizen journalism

W4RA, an international networked community set up by the Network Institute of VU University Amsterdam, the World Wide Web Foundation and Sahel Eco is about to start a new project: Foroba Blon. In Foroba Blon the W4RA team will develop innovative voice – web based services that support community radio stations in Mali by providing better ways to exchange and access news and information.

The project is funded by the International Press Institute (IPI), an European organization aimed at helping developing countries establishing and strengthening independent media organisations to support freedom of expression and freedom of access to information.

Foroba Blon will provide a voice-based information platform for people from rural communities in Mali – who currently have community radio as their only source of information. The platform will enable information exchange using voice  and will be accessible through simple, old-fashioned mobile phones.The platform will also have a web interface.

Foroba Blon may give a voice to people who were always underrepresented in the voices heard on the radio: e.g. women, disabled people, mobile pastoralists.

The Web of Speech

The World Wide Web is for us all, but do we all benefit from it?

Currently, 4.5 billion people in the world do not have access to the Web and all its advantages for publishing, sharing and searching information. Local conditions in developing countries still pose enormous challenges in terms of access and availability of relevant and usable web content. If we take a look at rural communities in West-Africa, communication is mainly voice-based – many people don’t read or write. People may have mobile phones, but these are simply used for talking. Even the SMS function is hardly used.

Great benefits are being attributed to indigenous knowledge sharing. In remote rural areas of the Sahel, subsistence farmers learn from each other how to “regreen” their fields and improve their crops and livelihood. Unsurprisingly, for these farmers the current Web of text and images, and most prominently, the Web for internet-users, is still not their most appropriate platform for knowledge sharing. Unless people are able to create their own content and publish it and share it with each other, the Web does not bring them an added value. Therefore, we must try to add voice functionality to the Web, and make it accessible using simple mobile phones. We must address the fact that there are many different languages spoken in this part of Africa. Languages that have never been written down on paper…

Fortunately, very recently, new research projects have started in support of rural communities in Mali, Ghana and Burkina Faso. This will enable people here to create their own “voice Web content”. This will soon result in the creation of voice web-content, such as community radio programs and voice messages from farmers.

As we all know, spoken text is actually a very different medium than written text. The new challenge will be to make spoken information searchable and retrievable, and enable it to be filtered and presented through voice. If we really want to create a multimodal web with voice content in local African languages,if we want to invent the “Web of Speech”, we will need ways to index and retrieve speech content in user-friendly, reusable and scalable ways. Can we make use of local knowledge and develop methods for social tagging and self-tagging usable e.g. for the rural communities we are working with in the Sahel countries? Are there other ways to annotate voice content, without having to translate this content into written text…?

To be continued…..Mobile telephony in Africa "Zain"