Knowledge Discovery in Databases (KDD): The Data Science Lifecycle
Date:
In May 2026, I led an online workshop for University of Brasília researchers and students focused on the KDD (Knowledge Discovery in Databases) process. As someone working at the intersection of Software Engineering (at the Federal Court of Accounts - TCU) and AI Research (at Petrobras/LAMFO), my goal was to show how we can move beyond simple data collection to identify valid, novel, and potentially useful patterns in complex datasets.
The session was a deep dive into the end-to-end KDD pipeline—moving from data cleaning and preprocessing to data mining and the final interpretation of results. We discussed how this methodology is essential for transforming massive volumes of information into strategic knowledge that can drive decisions in both government and industry sectors.
Key highlights:
- A practical breakdown of the KDD stages: Selection, Preprocessing, Transformation, Data Mining, and Evaluation.
- Insights from my experience applying these concepts to real-world challenges in Finance and Oil & Gas.
- Bridging the gap between academic theory and the pragmatic needs of modern software and data engineering.