← All work
PreviewData Science · Data Investigation

DataTrace

A data investigation workbench for profiling datasets, detecting unusual observations, comparing groups, and turning statistical evidence into structured findings.

Project demo

01 — Problem

Real-world tabular datasets rarely arrive clean and ready for analysis. Missing values, duplicates, unusual observations, skewed distributions, and differences between groups can require several different analytical techniques to investigate.

02 — Solution

DataTrace combines data profiling, exploratory analysis, anomaly detection, statistical testing, and evidence-based findings into a single investigation workflow.

03 — Analysis

The application computes statistical and analytical results in the backend, allowing users to inspect distributions, relationships, unusual observations, and group differences without relying on an AI model to perform the underlying calculations.

04 — Evidence

Investigations can be turned into structured findings containing the observed result, supporting evidence, methodology, and relevant limitations. AI-generated explanations are grounded in the computed evidence and remain subject to human validation.

Contact

Have an idea you'd like to turn into a working product?

Let's discuss it.