Bag of Words Model
A bag of words model is essentially counting words per document.
Data Structure
Rows are documents and columns are words or phrases. Inevitably this is a sparse matrix.
Implementation
First the words and phrases across all documents must be tokenized.
Next, the count of each key word/phase is extracted from each document.
Finally, the output matrix must be interpreted. Semantic meaning is lost so care must be taken to ensure that sourced documents have a common context.