Advanced Analytics Patterns

Quick Answer: Power BI supports advanced analytics beyond standard charts. What-if parameters enable scenario analysis with sliders. Anomaly detection highlights unexpected data points. R/Python visuals run custom statistical code. Combining these techniques with AI visuals creates a comprehensive analytics layer.

What-If Parameters

What-if parameters create dynamic sliders for scenario analysis:

Modeling tab → New Parameter → What If
→ Configure: Name, Data type, Minimum, Maximum, Default, Increment

How It Works

A what-if parameter creates:

1. A new table with values from min to max
2. A measure returning the selected value
3. A slicer for user interaction

Example: Revenue Scenario

// What-if parameter: Growth Rate (0% to 50%, default 10%)
Projected Revenue =
SUM(Sales[Amount]) * (1 + [Growth Rate Value])

Users slide the Growth Rate slider to see how different growth assumptions affect projected revenue.

Common What-If Scenarios

• Price sensitivity — How does a 10% price increase affect revenue?
• Discount impact — What happens if we offer a 15% discount?
• Headcount planning — How many new hires are needed at different growth rates?
• Budget allocation — What is the impact of shifting budget between channels?

Anomaly Detection

For time-series line charts:

Analytics pane → Find Anomalies → Turn On
→ Configure sensitivity (higher = more anomalies detected)

What Anomaly Detection Does

1. Builds a statistical model of the expected data pattern
2. Identifies points that fall outside the expected range
3. Marks anomalies with special indicators on the chart
4. Provides explanations for each anomaly (which dimensions contributed)

Anomaly Explanations

Click an anomaly marker to see:

• The expected value vs. actual value
• Which categories deviated most from expectations
• Contributing factors ranked by impact
• Confidence level of the anomaly detection

R and Python Visuals

R Visual

Visualizations → R Script Visual → Add fields → Write R code

# Example: Correlation plot
library(ggplot2)
ggplot(dataset, aes(x=Revenue, y=Profit)) +
  geom_point() +
  geom_smooth(method="lm")

Python Visual

Visualizations → Python Visual → Add fields → Write Python code

# Example: Distribution plot
import matplotlib.pyplot as plt
import seaborn as sns
sns.histplot(dataset['Revenue'], kde=True)
plt.show()

Limitations of Script Visuals

• Render as images (not interactive)
• Require R/Python runtime installed
• Limited to 150,000 rows of input data
• May not work in the Power BI Service without Premium/Fabric

Combining Analytical Techniques

A comprehensive analytics page might include:

On the Exam

The PL-300 frequently tests:

• Creating and using what-if parameters for scenario analysis
• Configuring anomaly detection on line charts
• Understanding the limitations of R/Python visuals
• Combining multiple analytical features for comprehensive analysis