The FINRA department of market regulation is involved in identifying fraud by surveilling 99 percent of the equity markets and approximately 70 percent of the options markets. FINRA needed to rebuild a picture of the U.S. market by obtaining a list of all daily transactions from all U.S. stock exchanges.

In this eBook, learn how FINRA built a picture of the U.S. market using data pipelines that could scale to support the volumes of data they were ingesting— more than 100 billion events per day. Learn how FINRA:

• Built a unified analytics platform that democratized data and brought previously siloed teams together.
• Developed a Feature Framework that helped them create modular code that could be changed or modified anytime. Also, human feedback and additional features were easily incorporated to improve the models over time.
• Built interpretable and transparent features and machine learning models which showed the decision making process for fraud.