license: apache-2.0
task_categories:
- text-generation
language:
- en
size_categories:
- 1K<n<10K
Dataset Card for Cleaned CBC Curriculum Text Data
This dataset contains cleaned and normalized textual data about Kenya's Competency-Based Curriculum (CBC), sourced from various online platforms that publish curriculum-related content. It has been processed using a powerful custom Python script that includes HTML tag removal, sentence reconstruction, and optional enhancement using open-source LLMs.
This dataset card aims to provide transparency about the dataset's creation, use cases, and structure.
Dataset Details
Dataset Description
This dataset consists of cleaned and structured text segments from publicly available websites related to Kenya's Competency-Based Curriculum (CBC). The texts highlight CBC values, competencies, structure, and guiding principles.
All raw data was collected from online sources, then thoroughly cleaned to:
Remove noise like HTML tags and escape characters.
Fix broken or incomplete sentences.
Preserve meaningful educational content for NLP tasks.
Curated by: Kariuki James
Funded by: Not funded
Shared by: Kariuki James
Language(s) (NLP): English
License: Apache 2.0
Dataset Sources
- Data Sources: Educational websites publishing "" content in Kenya
- Paper []: N/A
- Demo []: N/A
Uses
Direct Use
This dataset is ideal for:
- Training NLP models to understand Kenyan education policies.
- Fine-tuning LLMs for education-related tasks.
- Text summarization and classification.
- Semantic search and question answering about CBC.
Out-of-Scope Use
- Misrepresentation or misuse of Kenyan education content.
- Using the dataset for commercial purposes without proper citation or validation.
Dataset Structure
Each line in the dataset is stored in JSONL format and follows the structure:
{"text": "The CBC intends to be more practical and student-based."}