Big Data: Distributed Data Management and Scalable Analytics

The course is subdivided into 2 parts: Big Data Management and Big Data Analytics. The part on Big Data Analytics builds on concepts introduced in the part Big Data Management.

Part I: Big Data Management:

1. Introduction & Map/Reduce

2. Spark

3. Streaming: Spark Streaming & Storm; Big Data Architectures

4. Consistency and Availability

5. Distributed and parallel query evaluation

6. Stream Processing and Sublinear Algorithms

Part II: Big Data Analytics:

1. Introduction

2. Batch distributed machine learning

3. Sequential machine learning and streaming

4. Recommender systems and Collaborative filtering

5. Deep learning

This is an introductory course on big data management and analytics. Its objective is to introduce students to the fundamental notions, principles, and research results concerning modern, scalable, and fault-tolerant ways for managing and analyzing massive amounts of data using parallel and distributed systems. Armed with this knowledge, the student will be able to decipher, use, and compare the plethora of big-data technologies currently used in industry.

Learning outcomes

After successful completion of this course, the student:

1. Understands the characteristics of big data, and the challenges these represent
2. Knows the principal architectures of Big Data Management and Analytics Systems (BDMAS), is able to explain the purpose of each their components, and is able to recognize and explain the key properties, strengths and limitations of each type of BDMAS and their components.
3. Understands the key bottlenecks in managing and analyzing massive amounts of data and is familiar with modern algorithms for overcoming these bottlenecks using parallel and distributed computation.
4. Is able to actively use this algorithmic knowledge in the design and implementation of applications that solve common data management and analytics problems using different types of BDMAS.
5. Is able to build applications using specific instances of each type of BDMAS.
6. In addition, is able to use established software frameworks for reproducing/sharing her/his results, including virtualization software (Docker), version control systems (Git), and notebooks (Jupyter, Zeppelin)
7. Is able to implement an analytics pipeline (e.g. in Spark) able to process and learn predictive models from massive datasets
    

Required and Corequired knowledge and skills

• Databases, SQL
• Supervised machine learning (classification, regression, feature selection)
• Basic notions of statistics and probability
• Programming in Python

Combination of Ex-Cathedra Lectures, Exercise sessions, Computer labs, Self-study, and Project Work.

Contribution to the teaching profile

• Be capable of formulating and solving complex or open-ended technical and scientific problems by using abstraction, modeling, simulation, and multi-disciplinary analysis while satisfying the requirements of university-level research and responding to requirements, constraints, the set context and the technical, socio-economical ethical and environmental stakes—all with the purpose of obtaining concrete solutions.

• Have in-depth knowledge and understanding of a structured body of knowledge, both transversal and specialised. Be capable of autonomously and critically following current trends and advances in this body of knowledge.

• Define, plan, manage, and execute projects taking into account their objectives, the available resources and constraints; assuring the coherence and quality of the work and deliverables.

• Work efficiently with other professionals (in group, in partnership, or in competition), make decisions and develop leadership, in a variety of professional contexts, disciplines, and cultures.

• Communicate and share information in a structured manner: orally, graphically and written, in French and in one or more other languages. Communicate on scientific, technical and cultural aspects, adapting him/herself to the desired goal as well as the target audience.

References, bibliography, and recommended reading

See recommended references in the UV page.

Course notes

• Université virtuelle

Additional information

This is an advanced course: notions of relational databases, machine learning, and programming are required! In particular, the project requires that the student has a high degree of autonomy in programming and installing the required tools including virtualization software (Docker), version control systems (Git), and notebooks (Jupyter).
 

Contacts

Part I: Big Data Management: Pr. D. Sacharidis (ULB) 

Part II: Big Data Analytics: Pr. G. Bontempi (ULB) 

Campus

Plaine, Solbosch

Method(s) of evaluation

• written examination
• Project

written examination

Project

Combination of written exam and project work.

Mark calculation method (including weighting of intermediary marks)

Combination of written exam (8/20) and project work (12/20)

Language(s) of evaluation

• english
• (if applicable french )