Retail Analytics is the practice of leveraging all available datasets in the modern omnichannel environment (online and offline sales, baskets, advertising, promotions, prices, and product descriptions) to address real-world business challenges for retailers, marketplaces, and direct-to-consumer companies. This class addresses these challenges using modern statistical and machine learning methods, as well as data engineering pipelines:

1. Unsupervised Learning Creating and analyzing product embeddings from text and images (P2V-MAP, t-SNE, ViT Transformers) and conducting assortment analysis.  
2. Supervised Learning Time-series prediction using classical methods (ARIMA, Prophet) and comparing their predictions with modern deep learning methods (LSTM, N-BEATS). Building recommender engines using attribute- and embedding-based data inputs. 
3. Semi-Supervised Learning Using and applying named-entity recognition models to tag product attributes. 

Students will learn to predict product- and store-level sales, perform assortment analysis, and forecast demand for new product introductions. Emphasis is placed on mastering the end-to-end process of building applied data science research projects: From data extraction, transformation, and loading to algorithm development and minimum viable product thinking. 

Practical learning is reinforced through hands-on group projects using multiple authentic product/store/basket-level sales datasets. By the end of the course, students will be able to take analytical prototypes into production and address complex business problems with data-driven solutions.

1.  Understand omnichannel analytics measurement, modeling and management
2. Ability to clean, prepare and investigate large scale datasets consisting of products, assortment, advertising and other retail based data sets
3. Apply classical statistical methods on retail analytics datasets to solve marketing problems such as diffusion, sales prediction and product clustering.
4. Explore modern machine learning and data-driven product/store representation through text, image, and multi-modal learning.

Attendance is mandatory, including:

• participation during the sessions
• presentation of solutions
• final group presentations of all participants

• Classes will introduce algorithms and models and put them into a business context
•  Three datasets will be provided to all students to apply the learned concepts and present their results (same dataset for all students)
• Group projects with separate datasets/clients will be analyzed and presented by 3-5 students in the last class. Students select the most valuable use case from the lectures and apply accordingly.

Deliverables

Participation during the session (15%)

3 class assignments (45%)

1 Group Project (40%)

Grading

1: 90.0%-100%

2: 80.0%-89.99%

3: 70.0%-79.99%

4: 3: 60.0%-69.99%

5: <60%

Please log in with your WU account to use all functionalities of read!t. For off-campus access to our licensed electronic resources, remember to activate your VPN connection connection. In case you encounter any technical problems or have questions regarding read!t, please feel free to contact the library at readinglists@wu.ac.at.

Introduction Course to Statistical Learning or similar

During the weeks of the sessions: before and after class; for longer question, by appointment.