Senior Data Scientist - NielsenIQ (NIQ)
Text
Segmentation
ClassificationExtended product-classification models and partnered with operations to clean approximately 500,000 SPUs monthly, improving recall by 10% and precision by 5%. Improved O2O product categorization by 10-15% using BERT and TextCNN, and delivered an end-to-end reporting workflow covering approximately CNY 3.5B in FMCG GMV. Applied maximum-likelihood estimation to allocate SPU totals across approximately 1M SKUs, supporting measurement of billions of CNY in GMV. Built regression and sampling methods for unmeasured store segments (R2 approximately 0.90), improving reporting granularity by approximately 50% within budget constraints. Validated approximately 5M transaction-log records and produced 1,600+ client-facing market facts under NIQ quality standards, checking anomalies, methodology, and business interpretation.