SAS Reporting Migration: SAS Visual Analytics to Power BI¶

Audience: Report Developers, Business Analysts, SAS VA Administrators Purpose: Migrate SAS Visual Analytics reports, SAS Stored Processes, ODS output, and SAS/GRAPH visualizations to Power BI, Power BI paginated reports, and Fabric notebooks.

1. Overview¶

SAS Visual Analytics (VA) is SAS's self-service BI and data exploration tool. Power BI is the Azure-native replacement that integrates with Fabric lakehouses, Azure ML, and the broader Microsoft ecosystem. This migration is typically the lowest-risk, highest-ROI first step in a SAS-to-Azure journey because Power BI is a mature product with no code conversion required --- reports are rebuilt, not translated.

2. Migration approach¶

2.1 Report inventory¶

Before migrating, inventory all SAS VA reports:

For each SAS VA report, document:
- Report name and description
- Data source (SAS table, LASR table, CAS table)
- Number of pages/sections
- Visual types used (bar, line, scatter, map, table, etc.)
- Prompts/parameters
- Filters (global, section, visual)
- Calculated columns / derived items
- User count (consumers and editors)
- Refresh frequency
- Distribution method (email, web, mobile)
- Criticality (high / medium / low)

2.2 Priority framework¶

3. Visual object mapping¶

3.1 Standard visuals¶

3.2 SAS VA calculated items to DAX measures¶

SAS VA calculated item (profit margin):

Name: Profit_Margin
Expression: (Revenue - Cost) / Revenue
Format: PERCENT8.1

Power BI DAX measure:

Profit Margin =
DIVIDE(
    SUM('Sales'[Revenue]) - SUM('Sales'[Cost]),
    SUM('Sales'[Revenue]),
    0
)

SAS VA calculated item (year-over-year growth):

Name: YoY_Growth
Expression: (Current_Year_Revenue - Prior_Year_Revenue) / Prior_Year_Revenue

Power BI DAX measure:

YoY Growth =
VAR CurrentYear = SUM('Sales'[Revenue])
VAR PriorYear =
    CALCULATE(
        SUM('Sales'[Revenue]),
        SAMEPERIODLASTYEAR('Calendar'[Date])
    )
RETURN
    DIVIDE(CurrentYear - PriorYear, PriorYear, 0)

SAS VA running total:

Name: Running_Total
Expression: Running total of Revenue by Date

Power BI DAX:

Running Total =
CALCULATE(
    SUM('Sales'[Revenue]),
    FILTER(
        ALL('Calendar'[Date]),
        'Calendar'[Date] <= MAX('Calendar'[Date])
    )
)

4. SAS VA prompts to Power BI slicers¶

4.1 Dropdown prompt to slicer¶

SAS VA prompt:

Prompt Name: Select Region
Type: Drop-down list
Data Source: DISTINCT region FROM fact_sales
Default: All
Multi-select: Yes

Power BI:

1. Add a Slicer visual to the report page
2. Drag the Region field to the slicer
3. Format: Dropdown style, multi-select enabled
4. Set "Select all" option to enabled

4.2 Date range prompt to date slicer¶

SAS VA:

Prompt Name: Date Range
Type: Date range
Default: Current fiscal year

Power BI:

1. Add a Slicer visual
2. Drag Date field
3. Format: Between (date range slider)
4. Add relative date filtering: "is in the last 1 fiscal year"

4.3 Cascading prompts to synced slicers¶

SAS VA cascading prompts:

Prompt 1: Select State -> filters available Cities
Prompt 2: Select City (filtered by State)

Power BI: Cascading is automatic in Power BI. When State slicer filters the data model, the City slicer automatically shows only cities in selected states. No additional configuration needed --- this is a core Power BI behavior.

5. SAS Stored Processes to Power BI paginated reports¶

SAS Stored Processes are parameterized server-side programs that generate formatted output (HTML, PDF, Excel). The Power BI equivalent is Paginated Reports (based on SQL Server Reporting Services / SSRS engine).

5.1 Feature mapping¶

5.2 Migration steps¶

1. Identify the data query in the SAS Stored Process (the PROC SQL or DATA Step that generates data)
2. Recreate the query as a SQL query against the Fabric lakehouse or Azure SQL database
3. Build the report layout in Power BI Report Builder
4. Map parameters from SAS prompts to paginated report parameters
5. Configure subscriptions for scheduled email delivery
6. Publish to Power BI Premium workspace

6. Data source migration¶

6.1 SAS VA data sources to Power BI semantic models¶

6.2 Direct Lake semantic model (recommended)¶

Architecture:
  SAS Data --> Delta Table (Fabric Lakehouse) --> Direct Lake Semantic Model --> Power BI Report

Direct Lake provides:

• No data movement (reads Delta files directly from OneLake)
• Near-real-time refresh (as Delta tables update, reports reflect changes)
• VertiPaq performance with lakehouse-scale data
• Single source of truth shared with notebooks, Azure ML, and SAS (via SAS on Fabric)

7. SAS/GRAPH to Python visualization¶

For reports that include SAS/GRAPH output embedded in ODS documents, the equivalent in csa-inabox is matplotlib/seaborn/plotly in Fabric notebooks.

7.1 PROC SGPLOT to matplotlib¶

SAS:

proc sgplot data=work.quarterly;
  vbar quarter / response=revenue group=region groupdisplay=cluster;
  xaxis label="Quarter";
  yaxis label="Revenue ($M)" grid;
  keylegend / title="Region" position=topright;
  title "Quarterly Revenue by Region";
run;

Python (matplotlib):

import matplotlib.pyplot as plt
import numpy as np

fig, ax = plt.subplots(figsize=(12, 6))

quarters = df['quarter'].unique()
regions = df['region'].unique()
x = np.arange(len(quarters))
width = 0.8 / len(regions)

for i, region in enumerate(regions):
    region_data = df[df['region'] == region]
    values = [region_data[region_data['quarter'] == q]['revenue'].sum()
              for q in quarters]
    ax.bar(x + i * width, values, width, label=region)

ax.set_xlabel('Quarter')
ax.set_ylabel('Revenue ($M)')
ax.set_title('Quarterly Revenue by Region')
ax.set_xticks(x + width * (len(regions) - 1) / 2)
ax.set_xticklabels(quarters)
ax.legend(title='Region', loc='upper right')
ax.grid(axis='y', alpha=0.3)
plt.tight_layout()
plt.show()

7.2 PROC SGPANEL to seaborn FacetGrid¶

SAS:

proc sgpanel data=work.clinical;
  panelby treatment / columns=2;
  scatter x=baseline_score y=outcome / group=sex;
  reg x=baseline_score y=outcome / group=sex;
  rowaxis label="Outcome Score";
  colaxis label="Baseline Score";
run;

Python (seaborn):

import seaborn as sns

g = sns.FacetGrid(df, col='treatment', col_wrap=2, height=5)
g.map_dataframe(sns.scatterplot, x='baseline_score', y='outcome', hue='sex')
g.map_dataframe(sns.regplot, x='baseline_score', y='outcome',
                scatter=False, ci=None)
g.add_legend()
g.set_axis_labels('Baseline Score', 'Outcome Score')
plt.tight_layout()
plt.show()

8. Migration timeline¶

For a typical SAS VA estate with 30--80 reports:

9. Power BI advantages over SAS VA¶

10. Validation checklist¶

Maintainers: csa-inabox core team Last updated: 2026-04-30