United StatesFinance and insurance2024 – ongoingState Street
Deep learning model for data quality checks on investment data
State Street's deep learning model removed 87% of false data error alerts
Problem
State Street's data operations teams checked portfolio data with manual, threshold-based filters. The filters raised many false alerts that analysts had to review.
Approach
State Street built a deep learning model on Snowflake for data quality and anomaly detection. It flags portfolio market value issues for analysts to verify and replaces the fixed thresholds. The platform also supports natural language queries for risk analysts and portfolio managers.
Results
Takeaway
The model kept every real exception in view while it removed most of the noise. Data operations teams then spent their time on true errors.
Categories
- Size
- Large enterprise
- Industry
- Finance and insurance
- Business function
- Finance, risk and legalProduction and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Built in-house
- Evidence
- Vendor case study
Sources
Source summaries
Snowflake describes State Street's deep learning model for anomaly detection in investment data. It reports that the models detected all true exceptions, removed 87% of false positives and gave data operations teams a 25x productivity gain. The page carries no date; the Internet Archive first saved it on June 14, 2024.
Full details are in the original sources.