EP2420 Network Analytics 7.5 credits


This project course introduces students to data-driven engineering of networks and cloud systems. Using methods from statistical learning, students will develop and evaluate, for instance, models for prediction and forecasting of Key Performance Indicators (KPIs) and for anomaly detection. The models will be fitted and evaluated using testbed measurements or traces from operational systems. The functions built from these models are designed for real-time execution.

To develop the models, tools and packages from data science will be used, e.g., Jupyter notebook, scikit-learn, TensorFlow.

 The course is structured as two consecutive project blocks. Each block starts with introductory lectures that give background and discuss concepts for the specific project, followed by project execution, writing of a short report, and interview. 

  • Education cycle

    Second cycle
  • Main field of study

    Electrical Engineering
  • Grading scale

    A, B, C, D, E, FX, F

Course offerings

Autumn 19 for programme students

Autumn 18 for programme students CANCELLED

Intended learning outcomes

After passing this course, participants should be able to:

- perform the modeling of a network analytics task

- pre-process data and create predictive models using machine-learning techniques and tools

- assess, interpret and possibly apply the results 

- produce a written report describing and explaining the project results 

Course main content

This is a project course whereby students, by themselves or in small groups, perform an analytics project using data from a real system, for instance, using operational data from a network or compute cloud.

The course includes:

- introductory lectures on the specific machine-learning techniques used in the project

- an introduction into the tools to be used

- execution of the project by students, supported by message board and project meetings

- preparation of the project report by students

The specific project the students work on can change from year to year.


For single course students: 120 credits and documented proficiency in English B or equivalent

Recommended prerequisites

Basic knowledge in statistics, machine learning, networking, and computing systems. The projects require programming in Python.


Chapters from textbooks, online material, research papers will selected according to the specific project and will be made made available to students.

Required equipment

Students are assumed to have access to computers.


  • PRO1 - Project Work, 7.5, grading scale: A, B, C, D, E, FX, F

Grading will be based on the on the project report and the project interview.

The report and the interview will have equal weight for the grade.

In the project report, the project results will be evaluated in terms of correctness and completeness, and the presentation will be evaluated regarding structure and readability.

In the interview, students will be evaluated for their understanding of the project objective, approach, and results.

Grade A means the student has executed the complete project, has obtained correct results, and has produced a readable and concise report. Further, the student has answered well to all interview questions.

To pass the course, the student has executed the complete project, has obtained correct results for significant project parts, has produced a readable report, and has answered correctly to most interview questions. 

Requirements for final grade

Requirements for passing the course are that the student successfully completes both course projects and passes an assessment interview.

Offered by

EECS/Computer Science


Rolf Stadler <stadler@kth.se>

Supplementary information

Course documentation and background literature will be available through the course web site.

Replacing EP2400 Network Algorithms


Course syllabus valid from: Spring 2019.
Examination information valid from: Spring 2019.