Structured data is becoming critical in every domain and its availability on the web is increasing rapidly. Despite its abundance and variety of applications, we know very little about how people find data, understand it, and put it to use. This work aims to inform the design of data discovery tools and technologies from a user centred perspective, aiming to better understand how we can support people in finding and selecting data that is useful for their tasks. We approached this by advancing our understanding of user behaviour in structured-data discovery through a mixed-methods study looking at the work flow of data practitioners when searching for data. From that we present a framework for structured data interaction describing data-centric tasks, search strategies, as well as an in-depth characterisation of selection criteria in data search. We identified textual summaries as a main element that supports the decision making process in information seeking activities for data. Based on these results we conducted a mixed-methods study to identify attributes that people consider important when describing a dataset. This enabled us to better define criteria for textual summaries of datasets for human consumption. We designed a set of template questions to help guide the summary writing process and conducted an online study to validate the applicability of dataset summaries in a dataset selection scenario. The findings of this work revealed unique interaction characteristics in information seeking for structured data. Our contributions can inform the design of data discovery tools, support the assessment of datasets and help make the exploration of structured data easier for a wide range of users.