Case Study: Kasikorn Bank achieves improved prediction accuracy using 35 years of data with Cloudera

Kasikorn Bank improving prediction accuracy with 35 years of data

Kasikorn Bank’s technology arm, KBTG, needed to move beyond legacy BI to build the digital bank of the future: accurately predict loan defaults and customer credit demand using long histories and diverse data. Their existing Netezza data warehouse limited analysis to roughly two years of pre-aggregated data and forced analysts to work with subsets, preventing scalable, self-service machine learning and real-time insights.

KBTG implemented Cloudera’s modern data platform to consolidate transactions and external data (including social media), currently storing 12+ years of data with plans for nearly 35 years and ingesting about seven million transactions daily. The platform delivers ~10x more storage and compute, cuts statement access from up to a week to seconds (estimated US$5M savings in printing/postal over five years), improves fraud detection with fewer false positives, and speeds loan approvals and targeted automated lending with flexible on-prem/cloud deployment.

Kasikorn Bank

Tul Roteseree

Deputy Managing Director