Visualization in Power BI and AI is a course in using Power BI's AI visuals, such as Key Influencers, the decomposition tree and anomaly detection, to explain what drives the numbers, then extending reports with Python and scikit-learn for clustering and prediction. It suits data analysts, BI developers and data scientists who already build reports in Power BI.
A good dashboard does more than show what sales were. It should help answer why the numbers changed, which customers are about to leave and which points are unusual enough to need attention now. Power BI has built-in AI visuals that answer these questions without code, and when the questions get harder you can extend the same report with Python and machine learning. Many users, however, do not know which features work with the licence they already have, or how to read AI results without misinterpreting them.
This course is built around Power BI Desktop. Day one goes deep into the AI visuals that need no Fabric capacity, namely Key Influencers, the decomposition tree, anomaly detection, forecasting and smart narrative, with the statistics needed to read them correctly, and explains the conditions for Copilot and the retirement of Q&A. Day two brings Python into Power Query and Python visuals, using pandas and scikit-learn to cluster customers, detect anomalies and build predictive models, then shows the results in a dashboard. It ends with a workshop that analyses customer data from start to finish. (2 days, 6 hours per day, 12 hours in total, Intermediate level.)
Course Details
This course is the 2026 edition of Visualization in Power BI and AI. It runs for 2 days, 6 hours per day (12 hours in total, 09:00-16:00), as lectures with labs. Intermediate level. Day one goes deep into the Power BI AI visuals that need no Fabric capacity. Day two uses Python, pandas and scikit-learn inside Power BI for clustering, anomaly detection and predictive models, then shows the results in a dashboard. The course focuses on analysis with AI visuals and machine learning; it does not cover Power BI fundamentals or AI agents such as Claude Code, which have their own courses at the institute. Learners take home the Power BI files, Python scripts and sample datasets for every lab.
Section 1: The AI Landscape in Power BI in 2026
Section 2: Lab: Key Influencers
Section 3: Lab: The Decomposition Tree
Section 4: Lab: Anomaly Detection and Forecasting
Section 5: Smart Narrative and Copilot
Section 6: Workshop: A Driver Analysis Dashboard with AI Visuals
Section 7: Lab: Setting Up Python for Power BI
Section 8: Lab: Python Visuals
Section 9: Lab: Clustering and Anomaly Detection
Section 10: Lab: Classification and Regression
Section 11: Workshop: Customer Analysis with AI, and Publishing
No public rounds are open right now. Join the waiting list and we will contact you first when the next round opens, or ask us on LINE. Or call 02-570-8449 or 088-807-9770
Key Influencers is an AI visual in Power BI that finds which factors most affect a value you care about, such as what leads to low satisfaction scores or higher sales. Put the field to analyse in Analyze and the possible factors in Explain by, and the visual ranks the factors and shows how much each one changes the outcome. The Top segments tab shows the groups where that outcome is most likely. The results show association rather than cause and effect, so check them before acting.
Install Python, create a virtual environment and install pandas, matplotlib and scikit-learn with pip. In Power BI Desktop, open Options, choose Python scripting and point the Python home directory at that environment. Then use Run Python script in Power Query, which passes the table in as the dataset variable, process it, for example by clustering customers, and return the results as new columns. To refresh such a report after publishing to the Power BI Service, you need a personal gateway on a machine with Python installed.
No. AI visuals such as Key Influencers, the decomposition tree, anomaly detection, forecasting and smart narrative work in the free Power BI Desktop, and Python with scikit-learn is free as well. Copilot in Power BI needs Fabric capacity of F2 or above, or Power BI Premium, enabled by an administrator, so the course covers Copilot as an overview and demonstration and every learner can still complete the labs.
It suits data analysts, BI developers, data scientists and digital marketers who already use Power BI and want deeper analysis with AI. You should have built reports in Power BI Desktop, understand data models, write basic DAX and know basic statistics. You do not need to be a Python programmer, but you should be ready to read and edit short code samples.
You will be able to analyse data with Key Influencers, the decomposition tree, anomaly detection, forecasting and smart narrative and read the results correctly, set up Python in Power BI, use pandas and scikit-learn to cluster customers, detect anomalies and build predictive models, and show the results in a dashboard with a refresh plan for publishing. Participants who complete the training receive a Certificate of Completion with its own number, which can be verified online.
Yes. It can be run in-house for your team, with examples adapted to your own work, such as sales or customer data in the same shape your team actually analyses. You can request a quotation from this page, or contact us on LINE @itgenius or by phone on 02-570-8449.
Most organisations already have data on sales, customers, stock and marketing, but it is spread across many files and systems, Excel reports have to be rebuilt by hand every month, and the charts often miss the business question. Tableau Desktop lets you…
Many organisations keep their data spread across Excel workbooks, CSV files and several databases. Every month end, staff copy, paste and reformat the same reports, figures from different departments do not match, and managers do not see the big picture in…
Many organisations keep sales, budget and register data in Excel, but when they try to report on it in Power BI they find several subjects crammed into one sheet, merged cells, numbers stored as text, dates mixing Buddhist and Gregorian years, and fixes that…
