Contextual preferences to personalise semantic data lake exploration

In the latest years, the availability of data collected within Smart Cities is enabling citizens to take decisions about their daily life in an autonomous way. In this landscape, data aggregation according to different analysis dimensions may help users to take decisions, leveraging indicators as powerful tools for meaningful exploration. However, due to the volume and heterogeneity of Smart City data, data lakes have to be used as flexible repositories for enabling data storage and organisation. Despite they are usually based on centralisation of data storage, data lakes compel to consider pay-as-you-go or on-demand solutions, where integration is progressively carried out, to cope with the cumbersome nature of Big Data. Given the variety of interested users, their goals and preferences on available data, personalised data access, as well as representation and use of preferences, are required and need to be adapted to the unique characteristics of data lakes. In this paper, we describe an approach to model preferences on Smart City indicators built on top of a data lake. Preferences are used for personalised data exploration. Main contributions of this paper concern: (a) the definition of users’ preferences and preference constructors over the semantic representation of indicators; (b) the definition of users’ contexts and contextual preferences; (c) preference-based personalised exploration of Smart City data.