Evaluation · Beginner · 15 min
Reading a Quality Report
EvaluationQuality
Learn to interpret the evaluation report the way a skeptical reviewer would.
Steps
Follow along
Each step is something you can do in the platform right now.
01
Step 1
Run Evaluation on the synthetic dataset from the first tutorial.
02
Step 2
Look at the Overall Quality Score first — it summarizes everything.
03
Step 3
Check Correlation Similarity: are the relationships between columns preserved?
04
Step 4
Scan the Distribution Match scores per column for weak spots.
05
Step 5
Open the Linear Map projection and look for systematic divergence between real and synthetic.
06
Step 6
Check the Privacy Shield numbers before sharing the dataset outside your lab.
Notes
Pro tips
What experienced users wish they knew the first time.
No single metric decides quality — the batch is trustworthy when quality, efficacy and privacy all point the same way.
If scores are weak, try a different data synthesis mode or fewer generated rows before regenerating.
Want the details? This tutorial pairs with the docs page metrics · evaluation metrics.
Read the docs →