Data-to-text is an active research topic in the field of natural language generation (NLG) studying computational techniques for discovering and telling stories from data. Data-to-text NLG technology is being commercially developed and deployed in a broad range of industry verticals from weather to finance and from engineering to medicine. An important feature of these NLG systems is that they bring together data science and NLG. This talk presents a few case studies from academia and industry to explore the challenges and opportunities in studying and building data-to-text NLG systems.