Data Management

Data Management 2023/2024

The Data Management course is held from February to May 2024. It is a 6 ECTS course for students enrolled in the Master of Science in Engineering in Computer Science or in the Laurea Magistrale in Ingegneria Gestionale at Sapienza - University of Rome. The main instructor of the course is prof. Maurizio Lenzerini.

The course aims at providing a good knowledge of the structure and the functionalities of both relational Database Management systems (DBMSs) and NoSQL data models (including a hands-on approach).

For a detailed description of the syllabus, please refer to the link below. 

Website of the course

Exam

In order to pass the Data Management exam, two different options exist:

1. Do just the full written exam (all problems are mandatory). The evaluation will follow the standard rules, i.e., the minimum grade to pass the exam is 18 and the maximum grade is 30.

2. Do both a shortened written exam and a practical project on the topics available (Data Warehousing, NoSQL). In this case the minimum grade to pass the written test is 15 and the maximum grade is 24. The minimum grade to pass the project is 4 and the maximum grade is 8. Provided that the scores of both parts are sufficient, the final grade is given as the sum of the two evaluations. This options gives the opportunity to get a minimum grade of 19 and a maximum grade of 32 (30L).
   
   Further details can be found in this document.

F.A.Q.

Q: Are there specific dates devoted to presentations of projects?
A: No. As soon as you complete your project, just ask me for an appointment following the indications provided in the slides. 

Q: Do project presentations have to be done in presence?
A: No. They can be done remotely at the Google Meet address indicated above.

Q: My project includes code and I would like to share it as part of the project evaluation. How can I do that?
A: To share the code of your project, please provide a link to the corresponding repository.

Q: I already passed the written test but not the project presentation. What is the last date for presenting my project in order to have my final grade registered within a specific exam session?
A: I don't know.

Q: I already passed the written test but not the project presentation. Do I have to book for the written exam again on Infostud to have my final grade registered?
A: Yes. If you are about to complete your project and plan to register the final grade within an upcoming exam session, I suggest you to book for that session and inform prof. Lenzerini that you have booked for registration purposes only. If you have your project presentation outside of any exam session, you will be able to book for the first useful exam session (informing prof. Lenzerini).

Q: I am not satisfied with the grade obtained at the project presentation. Can I make some changes and present it again to have my grade reviewed?
A: No. If you refuse your grade you are required to go with a brand new project.

Q: I did my project presentation in A.Y. 2022/2023 and I got it evaluated out of 6 points. Can my grade be adapted to the new rating system (8 points)?
A: No. If you want to adopt the new exam modality you have to do a new project. Notice that there is no direct correspondence between 6 points obtained in A.Y. 22/23 and 8 points obtained in A.Y. 23/24. The new higher scores are equivalent to an honors grade for particularly meritorious projects .

Q: I passed the written test in A.Y. 2022/2023. Can my project be evaluated out of 8 points instead of 6?
A: No, unless you take the written test again. As for students who already passed the project presentation but not the written test, if you want to adopt the new modality you are required to take any exam part again.

Q: Which are the requirements to obtain the maximum grade for the project part?  
A: In order for a project to be evaluated with the highest score, it must meet certain criteria, including: demonstrating a reasonable level of complexity, providing a non-trivial and accurate modeling, and including detailed and well-motivated analyses. Notice that questions can be posed to students to establish whether or not to assign an honors grade.

Q: I want to do a project on Data Warehousing. Which are the relevant aspects to consider to obtain a high grade? 
A: First and foremost, choose the dataset carefully. Try to do a sketch of the DFM schema of your Data Warehouse before making your project proposal. You are expected to provide a modeling which includes at least 3 non-trivial dimensions (e.g., they should contain some hierarchies to be used for OLAP operations). Show interesting insights over your Data Warehouse through OLAP sessions. Provide the SQL queries to be posed on the Star/Snowflake schema (you can choose which one to adopt, but you have to be able to motivate your choice). The work done for the ETL phase can contribute to the evaluation.

Q: I want to do a project based on the comparison between the relational model and a NoSQL-based model.  Which are the relevant aspects to consider to obtain a high grade?
A: Choose the dataset carefully. Too small datasets are to be avoided due to the risk of them not being suitable for the task of highlighting the peculiarities of different systems. Carefully design the conceptual modeling for both the relational case and the NoSQL case, justifying your choices and avoiding too trivial modeling (e.g., a graph representation with a single type of edge and/or node, or aggregates with a single collection of non-nested aggregate structures). Try to compare the two approaches on different aspects, e.g., performance on specific queries, complexity of the query languages, ease-of-use of the different systems, scalability, etc.

Q: I have found a dataset consisting of a set of csv files. Do I have to trat them as relational tables and import them entirely in my database?
A: No (unless you want to get a low grade). In most cases datasets online are not organized according to the relational model. They are just semi-structured texts containing data. It is up to you to understand the data, carefully design the proper conceptual model and load your data into the corresponding database. You are free to leave out data that is not relevant for your analysis and you are free (encouraged) to integrate different sources into a unified view (your database).