Skip to main content

5.3.18. Kenya

FLaReNet Summary

Most of the activities on LR in Kenya are being conducted at the University of Nairobi where several data and tools have been produced.

Contact Point Input

National/Regional contact: Wanjiku Ng'ang'a, University of Nairobi.

Programs

Most of the activities are being conducted at the University of Nairobi:

  - School of Computing & Informatics, University of Nairobi:

1. Development of OpenSource spell checkers, compatible with openoffice, using Hunspell tools, for Swahili, Gikuyu, Dholuo, Kamba.
2. Development of morphological analyzers using Xerox XFST for Swahili and Gikuyu.
3. Development of shallow, rule-based MT systems for Gikuyu-English.
4. Development of Swahili resource grammars for Grammatical Framework.
5. Development of automatic diacritic correction tool for Gikuyu, done as part of AnLoc's localization work.
6. Localization and customization of the Africa News Monitor, a variant of the JRC's European Media Monitor. This is at the early stages and involves the EMM team and UNDESA, Nairobi office.
7. Development of Online resources for Igbo - MRD, TTS, audio corpora of dialectal Igbo and its textual transcripts. Done in collaboration with the Igbo Archival Dictionary Project at Nnamdi Azikiwe University, Nigeria.
8. Corpora compilation - the SAWA corpus (Swahili), Gikuyu corpus, SMS corpus of Nairobi (contains English, Swahili and Nairobi Slang).
9. News monitoring and sentiment analysis for vernacular radio broadcasts - done in conjunction with Twaweza communications, a local firm dealing in media and communication.

  - Electrical Engineering Department, University of Nairobi:

Text to speech system for Swahili and Kenyan English. These voices are already licensed in some information services such as the National Farmer Information Service (NAFIS).

There is also work on Swahili Localization, as it can be found on Facebook, Twitter, OpenOffice and Microsoft services and products.