Abstract:

Forming high-quality data scientists is a complex activity because Data Science is a wide and complex area. In this paper we present our methodology and experiences for training future data analysts and data scientists during six occurrences of an "i-semester" (where the "i" stands for innovation). Our methodology is based on agility principles of continuous reinforcement of technical principles and continuous evaluation and application of learnt concepts in a business challenge. An i-semester, was part of the pilot program aimed to change our institution's teaching-learning model (known as Tec21 model). During i-semester, students are assigned to work in only one capstone project, where students face a real problem (business challenge) that demands their immediate comprehension and abstraction abilities to apply topics learned in the classroom to solve the challenge. The syllabus for the integrated course combines the contents of six "traditional" courses: Statistical interpretation of data, Database and Multidimensional Models, Business Processes, Data Science principles and Ethics in the Professional Life. It is important to promote Ethics and integral behavior while working with Artificial Intelligence tools and data which may be treated as sensible, private and/or personal asset.

A key component in i-semester was the involvement of a business partner which will provide a "real-life" business challenge with real data. Our partner had to be open to continuously interacting with students in favor of a rich learning experience. Therefore, prior to the first iteration of i-semester we performed an extensive scouting with the local industry and through personal contacts to identify and convince such a business partner. Our efforts paid off with a worldwide lead company in the cement industry. The main role of our business partner was to provide real data from one of their business processes and allow students to analyze them with different approaches to eventually find some insights, correlation, or tendency in the data that the company can use to improve their business.

In six semesters, we impacted 116 undergrad students from IT, computer science and Industrial engineering. Every semester with a new group of students, we improve the process. Students worked in teams of four. The criteria for forming groups was based on a personality assessment and personal interviews looking for a mix of technical, business and social skills in each team. Applying Agile techniques, iteration and continuous practice, the results were better than we expected according to comments from our business partners and students once they finish their degree and started a professional life. During final presentations students from technical backgrounds were able to communicate statistical results in a business language easier for managers from the manufacturing client company to understand.

The paper addresses our agile methodology and experiences in course design, selection of contents, teaching complex topics and applying them to a real project, presenting results and evaluating students.

Keywords:

University-Industry Experiences, Data Science, Data Analytics, Enhance learning.