Dear all,
a few clarifications about the NS exam now that lectures have come to an end.
The exam consists of two parts:
- written test, with 12 multiple choice questions, and 2 open questions about the content teached during lectures. A reference example is available in the course website under Course slides and material
- oral presentation, using powerpont-like slides that contain the most relevant outcomes of your project, and considering, if you are in a group, approx 5 min each student + 5 mins for introducing the dataset/project (for a total of 10min for a single student, 12min for two, 20min for three, etc.). I expect that each group will send the presentation slides + the code (or link to the code) at least the day before the oral exam.
For IPs only, there is an additional (but simple) duty:
- you will also be required to contribute to the project report written by SNA students, by providing a brief description of the techniques you used in the analysis, plus, obviously, figures and plots that you implemented with your code
Exam dates are already available in uniweb for you to enrol. Note that we set a unique day (13 Feb 2026) for the presentation of the interdisciplinary projects, which will be a joint presentation from SNA+NS students. We will provide, for this IP day, a detailed presentation schedule later on at the end of January.
The final score will be weighted 40% on the written and 60% on the oral part. Important aspect to take into account in your project (these will build the final score) are:
- dataset choice, i.e., the way you selcted your data, and the availability in the dataset of meaningful information that allows to interpret the analytical resuls (any network of nodes and edges, with no additional info is not considered a good choice)
- dataset description through simple analytics: try describing the data you play with according to what is meaningful in jour project
- chosen analytics, that should cover (but are not limited to) the aspects covered during lectures. Please make sure to correctly show/interpret your results, e.g., degree distributions are expected in the log form, any community/topic detection has to be tested against (at least) the four quality measures we identified during the course, etc.
- quality and meaningfulness of visualization, so for example it is expected that network plots are readable, with a decent layout, and readable text
- any implementation of new code that implements algorithms that were not covered in the course, for example extensions to a multilayer or temporal format, or different approaches for clustering on networks, etc., are highly welcome.
- clarity and efficacy in the oral presentation
I will be available through the whole month of January for any needed feedback.
I wish you all a Merry Christmas and a Happy New Year!
Best,
Tomaso Erseghe