Case Study: Edelman Data & Intelligence achieves scalable, high-quality training signals for brand trust models with Labelbox

Edelman Data & Intelligence builds production-grade trust models on tight timelines with Labelbox

Edelman Data & Intelligence (DxI) faced the challenge of building machine learning models to predict consumer trust in brands, but lacked scalable ways to collect and enrich text data from the web to create a high-quality training signal. They needed to incorporate domain expert judgment efficiently to reduce bias and overfitting in their proprietary trust algorithms.

Using Labelbox, Edelman DxI built a data-centric process to produce its training signal, folding in internal expert feedback to refine the data. The solution integrated with their AWS data lake via Labelbox's Python SDK, enabling programmatic project creation. As a result, the team trained multiple production-grade models on tight timelines and now has the infrastructure to continuously iterate on its training signal as projects scale.

Edelman Data & Intelligence (DxI)