Create a Calculation View with Differential Privacy in SAP Web IDE
- How to create Calculation View of type Cube using SAP Web IDE
- How to configure differential privacy to add noise to a column
Thomas JungNovember 21, 2022
Prerequisites
- This tutorial is designed for SAP HANA service for SAP Business Technology Platform. Consider newer tutorials designed for SAP HANA Cloud.
- You have completed the previous tutorials to load data into the table
TGT_SALARIES
>**This tutorial cannot be completed with a trial account.**
- Step 1
Create a new folder called models under
db/src. Create a new Calculation View and call itSALARIES_ANONYMIZED
Click on the
anonymizationnode and then click on the white canvas to drop it.
Use the
on the node to add a table as a data source. Choose
TGT_SALARIES
- Step 2
Double click on the
Anonymize_1node. This will open the mapping. Double click onTGT_SALARIESto add all of the columns to the output
Click Details and switch
k-anonymityto Differential Privacy
Configure the following parameters for anonymity
- Sequence Column: ID
- Epsilon: 0.1
- Sensitivity: 15000
- Noised Column: SALARY
For example:
For more information about these parameters check the SAP HANA Security Guide
- Step 3
Connect the
Anonymize_1node to the Aggregation node.
Double-click on the name of the node to move all the fields into the output columns.
Go into the Semantics node and switch
START_YEARto attribute
- Step 4
Save and Build the calculation view
Right-click on it and choose Data Preview
Click Raw Data and you will see the anonymized data:
See the negative salaries? The added noise guarantees privacy while keeping sum and average results similar to the original dataset.
Use the Analysis tab and drag the
GENDERandREGIONfields to the label axis andIDto the Value axis.
Change the aggregation forSALARYtoCOUNTto answer the question below.