quincywang

Led and actively participated in planning and constructing the banking business data warehouse. Responsible for parsing and building the banking business data warehouse, emphasizing detailed features related to banking operations. Successfully constructed a comprehensive feature library covering key areas such as customers, transactions, and accounts. Established data dashboards centered around the customer lifecycle, including customer acquisition, transaction analysis, and risk assessment, aiding business teams in swift decision-making. Created multiple business-themed and goal-oriented intermediate tables within the existing data warehouse environment to support self-service analysis. Provided crucial financial risk data for finance and risk management teams, contributing to the formulation of risk management strategies and asset allocation decisions. Leveraged Spark for real-time customer profile creation, with a focus on disconnected customers. Crafted profiles based on key factors and maintained regular updates.

Collecting pertinent financial data, operational history, and other relevant information for small and micro-enterprises. Conducting meticulous data analysis to discern aspects such as profitability, debt-servicing capacity, and liquidity of the businesses. Increased data collection efficiency, reduced data preparation time, achieving a 30% speedup in adding new small and micro-enterprise data sources weekly. Lowered non-performing loan rates through data analysis, achieving a 5% reduction in non-performing loan rates. Developing and applying risk assessment models to predict the probability of default for small and micro-enterprise loans. Continuously refining and adjusting models to align with changes in the market and industry. Enhanced model accuracy, reducing prediction errors for loan defaults to within 3%. Optimized models for market adaptability, achieving a 10% quarterly improvement in model adaptability to market fluctuations.

Addressed integration issues swiftly, minimizing system glitches and delays in loan process. Identified anomalies through backend log analysis. Established real-time alert mechanisms for SMS volume, account balance, loan discrepancies, and credit rule monitoring. Implemented automated daily reporting via Feishu. Adjusted anti-fraud strategies, addressing group fraud and unethical practices. Built a data analysis system from scratch using Python and Excel. Introduced Superset for analytics on collection, SMS effectiveness, and overdue exposure. Used Airflow for data scheduling and improved dependencies. Selected AWS S3 + Redshift for cost effectiveness. Extracted data from MySQL and MongoDB to S3, loaded into Redshift. Implemented data warehouse layering for enhanced usability. Utilized flink and hudi  with AWS for a streaming processing system. Analyzed credit reports, built credit features using sklear. Conducted text feature mining, achieving a 4% increase in KS value from customer SMS texts.

Unified Customer Data Consolidation: Developed SQL scripts that improved data integration speed, reducing consolidation time by 30%. Simplified Cleaning Procedures for Third-Party Credit Data: Streamlined cleaning data, particularly for third-party credit data, resulting in a 5% reduction in data cleaning errors. Establishment of a Unique Bank-wide ID: Successfully established a unique bank-wide identification system, improving identity matching accuracy and reducing identity confusion. Streamlining HiveSQL for Improved Query Efficiency: By optimizing HiveSQL, query efficiency increased by 40%, with the average query response time reduced to half of the original duration. Participated in discussions to optimize the star rating system. Analyzed the impact of schemes on customers and adjusted operational strategies. Formulated business calibers, conducted data surveys, and assessed feasibility in the current tech environment.