Working with unstructured data
In the previous section, we explored some of the most common tasks and processes that are conducted when handing text-based data. More often than not, you will find that the data you work with is generally not of a structured nature, or perhaps not of a digital nature. Take, for example, a company that has decided to move all printed documents to a digital state. Or perhaps a company that maintains a large repository of documents, none of which are structured or organized. For tasks such as these, we can rely on several AWS products to come to our rescue. We will explore two of the most useful NLP tools in the next few sections.
OCR using AWS Textract
In my opinion, one of the most useful tools available within AWS is an Optical Character Recognition (OCR) tool known as AWS Textract. The main idea behind this tool is to enable users to extract text, tables, and other useful items from images or static PDF documents using pre-built machine learning...