Entry requirements

120 credits in science/engineering including Scientific Computing I, Database Design I and a second course in computer programming (programming in Java and/or Python). Scientific Computing I may be replaced by Scientific Computing, Bridging Course, 5 credits, or Numerical Methods and Simulation 5 credits, or Scientific Computing and Calculus, 10 credits.

Learning outcomes

To pass, the student should be able to

• use state-of-the art software platforms for management and processing of massive scientific datasets.
• analyse the characteristics of a data-intensive scientific application and propose suitable strategies to handle the data analytics aspects of the application.
• implement software to address an application’s data analysis needs using the technology presented in the course.
• critically analyse, discuss and present solutions and implementations in writing and orally.

Content

How to develop scientific applications utilizing methods and key concepts of large scale data processing platforms. Distributed file systems and cloud containers such as OpenStack Swift. Batch data processing with MapReduce based infrastructures such as Hadoop. Effective use of query languages such as Hive for scientific applications. Effective use of indexing. Array databases such as SciDB, data stream processing platforms such as Storm. Overview of techniques, concepts and tools for analysing massive data, such as NoSQL, NoDB, flat namespaces and ontology based data.

Instruction

Lectures, guest lectures, laboratory work, seminars and group supervision.

Assessment

Active participation in seminars. Written and oral presentation of assignments, a software project and research papers.