DATA SCIENCE FOR BUSINESS
- Overview
- Assessment methods
- Learning objectives
- Contents
- Bibliography
- Delivery method
- Teaching methods
- Contacts/Info
The student knows about the topics presented in the courses of Intelligent Systems and Data Mining.
A knowledge of at least one programming language is helpful. Also, it is suggested to bring a laptop.
The exam is aimed at assessing the acquisition and correct understanding of the discipline and the main topics presented during the course.
For the exam all students will have to implement a project (individually or in small groups) suggested by the teachers and sustain an oral interview. The interview will start from the description of the project work.
The final grade is awarded in thirty and it is obtained taking into account both the project and the oral interview.
The evaluation criteria will be as follows: quality of the implementation of the project, ability summarize the acquired knowledge, ability to independently identify the theoretical elements to be used to solve the proposed problems and to develop a solution strategy, knowledge of the main difficulties and pitfalls to avoid in executing projects in the presented field. The final grade will also evaluate the communication ability, shown in exposing the answers and the ability to properly motivate statements, analysis and judgments.
The exam is succeeded with a minimum grade of 18/30.
The aim of this course is to complete the presentation of data science-related topics bent towards business applications. Starting from the topics presented in the Intelligence systems and Data Mining courses, this course will provide methods and techniques geared towards the implementation of projects suitable for production environments.
This course relates the knowledge of theoretical aspects of data science with the most relevant technologies for manipulating, managing and visualizing data.
Students will be taught about data analysis through datasets available online. Throughout the course activities, predictive experiments (machine learning and deep learning) will be shown in order to fulfill the requirements of the use cases.
The learning objectives and expected results of this course are:
To define an architecture that allows for the development of a data science project, with respect to data volume, velocity and availability, along with computing power and implementation and maintainability requirements.
Select adequate methods for solving the proposed problems, with machine learning and deep learning technologies.
Analyze, visualize and meaningfully interpret the obtained results, given the proposed solution methods.
Implement simple projects in order to gain hands-on experience on methods and techniques in data analysis.
1. How a data science project works
1.a. Tools and cloud
1.b. How a project works
1.c. Projects’ examples and use cases
2. Data transformation and load
2.a. Data manipulation
2.b. ETL concepts
2.c. Data quality
3. Data analysis
3.a. Feature selection and class unbalancing
3.b. Data analysis workflow
3.c. How to contextualize and perform effective analyses
4. Machine learning and deep learning
4.a. Classification
4.b. Clustering
4.c. Feed forward networks
4.d. Autoencoders and Word2Vec
5. Presentation
5.a. How to effectively present data
5.b. Data representation with Business Intelligence tools
The course has no official textbook.
However, the main topics covered in the course are clearly described in this text:
Aurélien Géron, Hands-On Machine Learning with Scikit-Learn, Keras & TensorFlow, O’Reilly
Frontal lessons consist in 40 hours of theoretical lessons, alternating theoretical lessons and practice lessons. 16 hours of exercises consist in individual presentations and projects on selected datasets.
The teacher receives by appointment, upon request by e-mail to andrea.biancini@uninsubria.it