ETM5800 Text analytics for business

Today, many organisations face massive amounts of unstructured textual data such as social media data, product, and service reviews, the information generated from company websites, and others. However, the phrase “Data is the new oil” is only valid if companies can harness valuable insights from extensive unstructured data and use this to guide them in making better decisions for the company.

This course will provide the analytical tools to extract information from unstructured business-related textual data, derive patterns and trends, cluster the data, make inferences, and finally communicate or make predictions about the data. The course introduces powerful text analytical techniques using relevant computer software to administer these techniques. The lessons will begin with motivations for exploring text, identifying text format types, and other principles governing text data. After that, there will be an introduction to various text analysis software and the use of software to extract, clean, and inspect documents. The analysis section will begin with descriptive statistics and visualisation of textual data, followed by opinion mining using sentiment analysis, analysing word frequency and documents using tf-idf and examining relationships between words using n-grams and correlations. The course will then demonstrate the use of unsupervised machine learning topics to categorise information and discover hidden semantic structures in text data. Examples of these techniques are such as cluster analysis, topic modeling, word embeddings, and document embeddings. You will also be exposed to document classification models and techniques to fit and evaluate them. Finally, the course will discuss text data application for prediction and social network analysis. All practice exercises for the different text analytical methods will infuse real-world business examples to equip you with tools to relate text analytics with the practical business scenario.