The ubiquity of digital data storage spaces and the ease of acquiring and generating new data, via smartphones, tablets, embedded cameras, stationary surveillance cameras, GPS-type tracking systems, a Vitale card, medical imaging systems, etc., allow everyone and public bodies to store a phenomenal amount of data, as evidenced by digital libraries loaded with photos and personal videos, information stored by social networks, or open data portals recently made available by the regions and the French state. Typically generated and stored data can be heterogeneous and varied: for example, they may relate to images of the lungs of a patient with a disease, to the shopping habits of customers of a store, or to the renewal of markings on the roads of the Eure department.The set of these data forms a wealth of wealth that it is necessary to explore, clean and interpret in order to generate knowledge. As also highlighted in the Mc Kinsey' 2011 report, this ability to extract knowledge from data will be in the very short term, and for a long time, a vehicle for innovation and wealth. As a result, scientific discoveries obtained through data analysis are increasing and are involved in several disciplinary fields such as biology, health, humanities and social sciences. ... In this context, data science is the central driving force of these research and innovations, and the scientific issues addressed therein address the problems of acquisition, storage, indexing, modelling and analysis. Among the difficult questions are the integration of these complex, heterogeneous and interdependent data in order to generate knowledge, assist in decision-making or generate value.The Haute-Normandie region and all institutional actors in the region, fully aware of the importance of the digital sector, multiply its development activities and support for local actors and entrepreneurs, generating or processing digital data; for example, the creation of Seine Innopolis and the digital canteen. In order to leverage these data, it is therefore of paramount importance to initiate a structuring of local activities related to data analysis and science. The project we propose here is a unique opportunity to launch and federate activities around this theme.The DAISI project aims to structure some of the work carried out in data science and involving various high-normand actors (see list of partners below). Its aim is to unite the strengths and means acquired and implemented by these actors in order to strengthen existing skills, to bring forth new fields of research born from the cross fertilisation of particular scientific fields and of data science, and finally to encourage actors to produce scientific results of international significance from their innovative aspects. From a scientific point of view, its objective is to develop innovative methods and tools to analyse large amounts of data, to extract information and knowledge useful for different disciplinary fields such as health, the urban environment and urban traffic analysis. From this point of view, DAISI is therefore an upstream research project guided by applications with high societal impact.