Demo — dataset analysis

What our engine detects on a training dataset, in seconds.

100% synthetic data (fictional credit scoring). No real data.

The dataset

510 rows · 10 columns

idBIGINT
ageBIGINT
sexVARCHAR
postal_codeBIGINT
incomeBIGINT
employment_yearsDOUBLE
loan_amountBIGINT
emailVARCHAR
full_nameVARCHAR
defaultedBIGINT

What the engine found

High Personal data (PII) column « email »
High Personal data (PII) column « full_name »
High Bias column « sex » ratio 2.1
Medium Duplicates 10 occurrence(s)
Medium Bias column « postal_code » ratio 1.8
Low Missing values column « income » 25 occurrence(s)
Low Outliers column « income » 3 occurrence(s)

Zoom: detected disparities

`sex`

Group	Count	« defaulted » rate
`F`	250	30 %
`M`	260	14 %

`postal_code`

Group	Count	« defaulted » rate
`93200`	100	31 %
`59000`	107	24 %
`13008`	103	19 %
`69003`	88	19 %
`75001`	112	17 %

The generated dossier

From this analysis, the engine produces an Article 10 dossier (skeleton, 8 blocks) ready to be completed by an expert.

📄 View the sample dossier

Want to see this on your own data?

Free assessment