Solutions for the Public Sector start with the ability to share data safely

globe figure

Ensure your mandated statistical products are rigorously protected against disclosure risk, while increasing the quality and utility of your data.

Request a Demo
arrow

use cases

Our solutions and platform enable agencies to release data more securely, and more frequently, while protecting individual privacy.

blue figure

our process

A collaborative, calibrated process to assure utility while maintaining privacy

Proven with data-driven leaders in the public sector, our process and platform deliver on your specific goals.

Define the problem
Outline the problem, explaining the data release purpose, strategy, privacy considerations, error metrics, and a pseudocode algorithm draft.
arrowdown arrow
Confirm the viability of using DP
Using default hyperparameters, see if it is possible to conduct a differentially- private data aggregation.
down arrow
Decide on error metrics to optimize
Develop internal error metrics to assess the utility of the differentially-private dataset, considering that while added noise is necessary for privacy, excessive noise can compromise data usefulness.
figure
down arrow
Experiment with a wide variety of hyperparameters
Before finalizing error metrics, conduct a grid search on hyperparameters such as output threshold, noise scale, and keyset to find the optimal set using the predefined error metrics.
arrowdown arrow
Productionize the pipeline
Finalize the script with integrated error calculation and privacy loss logging, then automate its execution to run at regular intervals.

Trusted by

Leading data-focused agencies work with Tumult

How we use differential privacy

Extract powerful data from sensitive data, while maintaining privacy

solutions-graphic
1
Raw personal data: too sensitive to use
2
Protective transformation, with DP
3
Safe to share summary data. DP offers valuable insights with guaranteed privacy.

Differential Privacy powers a responsible, defensible, future-proof approach to privacy.

Solutions for the public sector

Why is DP an ideal fit for the public sector’s data sharing use cases?

Differential privacy enhances privacy in public sector data sharing, supporting the improvement of public services and policy making without compromising individual privacy. It enables data analysis, policymaking, and research while protecting individual confidentiality.

This approach, mathematically proven to safeguard privacy, boosts public trust, crucial in times of increasing data breaches and privacy concerns. With assured privacy protection, individuals are more likely to share data, improving public dataset quality and accuracy.

Videos

Tumult Tune: 5-minute teaser demo

Watch a teaser demo of Tumult Tune, an upcoming product that allows you to easily understand and optimize differentially private data products.
Read more
View All

case studies

View All
Image of a collage website

Illuminating college outcomes, while protecting privacy

Public Sector

Joining sensitive data sets from the Department of Education and the IRS in a way that protected privacy resulted in College Scorecard - a platform that allows students and families to simultaneously consider the cost and evidenced outcomes of a range of possible degrees.

Read more
right arrow

Recent articles

View All
Research

AIMing Higher: A Smarter Approach to Privacy-Preserving Synthetic Data

Learn how the AIM algorithm, co-invented by Tumult Labs CEO Gerome Miklau, improves upon existing algorithms for synthetic data generation by adapting to the user’s analysis needs and capturing key patterns in the input data.

Research

A Winning Approach to Generating Synthetic Data

A scientific paper, co-authored by our CEO Gerome Miklau, introduces a cutting-edge method for generating differentially private synthetic data.

Blog Post

Publishing Wikipedia usage data with strong privacy guarantees

Tumult Labs helped engineers at the Wikimedia Foundation design, implement, and deploy a differentially private solution to publish Wikipedia usage metrics in a provably secure way.

Research

AIMing Higher: A Smarter Approach to Privacy-Preserving Synthetic Data

Learn how the AIM algorithm, co-invented by Tumult Labs CEO Gerome Miklau, improves upon existing algorithms for synthetic data generation by adapting to the user’s analysis needs and capturing key patterns in the input data.

Research

A Winning Approach to Generating Synthetic Data

A scientific paper, co-authored by our CEO Gerome Miklau, introduces a cutting-edge method for generating differentially private synthetic data.

Blog Post

Publishing Wikipedia usage data with strong privacy guarantees

Tumult Labs helped engineers at the Wikimedia Foundation design, implement, and deploy a differentially private solution to publish Wikipedia usage metrics in a provably secure way.

Unleash the power and value of your data.

Request a Demo
arrow