AI-Powered Business Intelligence

Chapter 9. Leveraging Unstructured Data with AI

In the previous chapters, we used AI quite a lot with structured data, or as most people call it, tables. However, a lot of data in businesses is not actually stored in clean tables, but comes in a plethora of formats such as PDFs, images, raw text, websites, and emails. When you consider these formats, the majority of data available within organizations is unstructured. With AI, we can unlock these treasure troves and get insights from data that has hardly been touched before by analysts or data that otherwise needs a lot of manual effort before anyone can get insights from it. In this chapter, we’ll explore how AI can help us analyze texts, documents, and image files.

Use Case: Getting Insights from Text Data

Written language is one of the biggest and most diverse data sources humanity has collected. And businesses are no exception. The biggest creators of data are people, either within or outside organizations. Customers become content producers and share their opinions about products or services across the web and on various channels.

In this use case, we are going to deploy an AI service that will help us make sense of this data. In the concrete problem at hand, we are going to analyze user reviews at scale and communicate key insights through a BI dashboard. Let’s go!

Problem Statement

Small rooms, unfriendly staff, and horrible breakfast—or not? The management of a large hotel is strained by the variety of customer feedback. ...