Mastercourse ‘Data Innovation beyond the Hype’

The mastercourse ‘Data Innovation beyond the Hype’ held by the Data and AI Competence Lab of Sirris (EluciDATA Lab) provides pragmatic and industry-oriented sessions on data-driven innovation. The mastercourse cycle consists of six general sessions, covering the different steps of the data science workflow. Additionally, the Lab has developed a set of new topic specific sessions focusing on industrial applications of data science and artificial intelligence.

General sessions

Session 1: The art of formulating a data science task

The goal of this session is to offer an overview of data science and its opportunities for innovation from an industrial point of view. By means of actual cases from current data innovation projects in several domains, the iterative and creative path from business understanding to data exploitation will be described by means of the different steps in the data science workflow. Particular attention will be given to the kind of challenges that can be tackled, and the data and skills that you need to have available in order to realise those challenges. Furthermore, some common beliefs about data analytics as a commodity are debunked.

Session 2: Bring your own data innovation challenge (on demand session)

Session 2a: Bring your own data innovation challenge: proposal 

In the first part of the session, participants are invited to bring a business problem and a description of their available data which could be used to solve this problem. Participants will first present their business challenge to the group and take part in an interactive discussion, to identify opportunities to leverage their data. Subsequently, working with the EluciDATA Lab experts, participants will develop a plan for analysing the data, in order to answer specific questions relevant to their use case. At the end of this session, participants will have a structured analysis plan.

Session 2b: Bring your own data innovation challenge: execution

In this second part of the session, with the guidance of the EluciDATA Lab experts, participants will carry out the analysis of their datasets according to the analysis plan they developed in the first session. An interactive computing environment using Jupyter notebooks and standard Python libraries will be made available, along with some additional custom tools designed to speed up analysis. At the end of this session, participants will have progressed towards extracting valuable insights from their data which could be applied towards their specific business challenges

Session 3: The importance of data exploration and hypothesis building

Having a clear idea of the business opportunity and a library of intelligent algorithms at your disposal is not sufficient for realizing a data-driven solution. A detailed understanding of the available data in order to derive viable working hypotheses about the underlying mechanisms of the problem under study is just as important. Is the data suitable to solve the business challenge? Is it of the right quality and nature? What data is missing? Does it exhibit significant patterns and trends that can be exploited to model and understand better the problem under study? In this session, several data exploration approaches will be discussed, with particular attention for numerical and visual data exploration techniques.

Session 4: The power of data visualisation

Data visualisation is a powerful mechanism, useful in several phases of the data science workflow: it supports data exploration and understanding, and enables to present insights extracted from this data to users. However, choosing the most effective data visualisation method is not straightforward. If not selected carefully, inappropriate visualisations might lead to incorrect interpretations. Choosing the most effective visualisation not only requires knowing which methods exist and are most suited, but also requires knowledge about the domain, knowledge about the problem, and knowledge about human perception. The goal of this session is to give an overview of the existing data visualisation methods and of how they can be most effectively used to highlight important data properties, emphasize trends, reveal hidden patterns, etc.

Session 5: The art of feature engineering

Any intelligent algorithm that is used to learn something from data requires that this data is presented in the most optimal way. The process of transforming the data and extracting the most relevant distinguishing characteristics out of it is called feature engineering. It is arguably the most important step in the data science workflow, as even the most intelligent algorithm will not produce satisfactory results if the used data does not capture the most essential properties of the phenomenon under study. There is no clearly-defined formal process for engineering features and consequently this requires a lot of creativity, iterations, domain knowledge, etc. The goal of this session is to give an overview of the most commonly used approaches, as well as lessons learnt and common pitfalls for different types of data (sensor data, location data, etc.) and problem settings (prediction, profiling, etc.).

Session 6: Choosing the right algorithm for the right task

The goal of this session is to introduce the participants to the most important data science tasks (classification, clustering, regression, etc.) and provide an overview of the most commonly used algorithms and techniques to solve each of these tasks. For each of the methods, its characteristics, advantages and disadvantages will be explained in order to guide the participants in making a conscious choice in terms of the available data (dimensionality, attribute types, etc.) and the expected model requirements (interpretability, accuracy, scalability, etc.). Finally, the guiding principles to train and evaluate the resulting models, including an overview of common pitfalls and frequently-used evaluation measures, will be presented.

Sessions focusing on industrial applications

Session on the opportunities and challenges of fleet-based analytics (27/01/2020) - Zwijnaarde

In this session, participants will learn about state-of-the art techniques for analysing fleets of assets, such as industrial machinery or vehicles. The goal of the session is to introduce techniques for solving common problems faced by fleet operators:

• identifying malfunctioning or underperforming assets
• dealing with lost data connections and missing data
• benchmarking assets against one another
• identifying remaining useful life of assets
• modelling new assets that lack historical data

In each case, the fleet context provides additional insights that can help you improve operational efficiency of your assets. Techniques are introduced at conceptual and intuitive level with practical real-world examples.

Session on the industrial relevance of advanced AI/ML methods (05/03/2020) - Heverlee

The goal of this session is to give participants a better insight into the type of problems that can be tackled using machine-learning techniques, illustrated by several real-world examples from a diverse set of domains. Among others, the following machine-learning methodologies will be covered:

• Reinforcement learning
• Deep learning
• Evolutionary algorithms
• Transfer learning

For each method, its characteristics, advantages and disadvantages will be explained in more detail, as well as the most commonly used algorithm(s) to solve a particular industrial problem. The aim is to guide the participants in making a conscious choice for the appropriate technique in function of the problem setting at hand, as well as the available data (dimensionality, attribute types, etc.) and the expected model requirements (interpretability, accuracy, scalability, etc.).

Practical information

Target audience

The mastercourse is open to any company that is interested in data innovation as an opportunity for its company and activities. For sessions 3 to 6, some basic analytic skills (e.g. high-level understanding of algebra and interpretation of statistical figures) are a prerequisite.

Pricing

 *The early bird price is only applicable if you choose to follow one session. Registration must be done at least more than 3 weeks in advance. 

**This price is applicable for any Sirris member who subscribes at any time to one single session. 

• All prices are exclusive of VAT.
• A hardcopy of the course notes is included in the registration price. All sessions are given in English.