The account visit feedback given by account and product managers of a
sales department is collected and saved into a database. Besides the
basic structure created by the storage in a database, the feedback is
free text which contains a lot of abbreviations and spoken language. The
concepts of text mining are used to create an editorial tool with the
goal to present the information to the analysts in a way that they are
able to aggregate it to useful and meaningful phrases. Principles like
text tokenization, lemmatization and categorization are explained.
Approaches like Document Clustering, Information Retrieval and
Information Extraction are consolidated into a general Text Mining
model. A practical implementation of this model, including some code
snippets in PHP and MySQL, leads to the implementation of an integrated
piece of software.