Ghana English ASR: 2,700 Hours of Transcribed News Broadcasts

Ghana NLP Community released this 2,700-hour collection of Ghanaian English speech and transcriptions in March 2026. Sourced from news media broadcasts, it contains up to 1,000,000 audio segments specifically for West African accent modeling.

Use Cases

• Fine-tuning ASR models using the audio and corrected_text columns
• Analyzing speech duration patterns using duration_ss
• Developing accent-specific speech-to-text systems for West African English

Strengths

• 2,700 hours of audio data
• Verbatim transcriptions in the corrected_text field
• Sourced from professional news media broadcasts

Limitations

• Vocabulary is likely biased toward news and formal broadcast topics
• Geographic representation is limited to Ghana

Provenance

Source

Ghanaian news media broadcasts

Collection Method

Extracted from broadcasts

Freshness

Last updated March 2026.

Geography

Ghana