AI Skills for Engineers: Data Engineering and Data Pipelines

Artificial Intelligence and Machine Learning have become central techniques for most services and products, ranging from web-based systems to medical procedures, self-driving cars – even intelligent coffee makers.

Alongside algorithms, data is central to AI applications. Without solid data management, AI projects typically underperform or even fail. Unfortunately, the relevance and complexity of handling data is frequently underestimated.

That’s why we developed this course which covers foundational questions like “Why is data important to AI?” and “What data does AI need?” and covers more application-oriented topics and skills like how to extract, load and query data using an SQL pipeline.

In the second part of the course, you will learn basic data engineering skills, including how to setup your Python notebook environment, explore data with advanced pandas functions, and create simple and clear data visualizations.

This introductory course is targeted at learners with little experience in data management or Python-based data management who want to develop Python-based AI applications in the future. The course covers a brief introduction into data management for AI, relational data management (e.g., SQL), and practical data handling skills in Python, pandas, and Jupyter.

This allows you to build a foundation to prepare for future AI and Machine Learning development with Python.

Week 1:

• We ask why we should care about data management for Artificial Intelligence and Machine Learning (ML) systems.
• We examine which data are needed in the ML lifecycle and what properties that data should have.
• We discuss the effort and time needed for data management activities, and look at possible data sources.

Week 2:

• The basic key concepts of data management, such as databases, data models and data schemas are all introduced.
• The Relational Data Model is explained and contrasted with the Single-Table Model (like CSV and Excel) and Document Models.

Week 3:

• We show how to extract data from existing relational databases using SQL queries and converting the query results into CSV files for further processing using pandas in Python notebooks.

Week 4:

• The different ways setoff setting up and running Python notebooks are covered, including cloud-based notebooks and local notebooks.
• We will take you step by step through the process of setting up your conda environment and installing Jupyter and pandas libraries.
• You will learn how to run notebooks in VS code.

Week 5:

• Become a pandas expert.
• Explore the essential functionalities of pandas and, most importantly, write elegant and efficient Python pandas code to process and engineer tabular data.

Week 6:

• You will learn how to make simple and clear scientific figures in Python using the Seaborn library.
• Use the core functions provided by Seaborn to make beautiful statistical plots.