Research Dataset Recommendation System
Matches research questions to datasets across 25+ sources; reduces search time from ~2 hours to <5 minutes.
~/portfolio $
Hi, I'm Luyang Si
// data-analytics · business-intelligence · behavioral-research · nlp
I build data pipelines, translate data into actionable business insights, conduct behavioral and product research, and apply NLP to real research problems.
$9 user behavior interviews + 4-round divergence study at RtB — translating decision patterns into measurable product signals for alpha-stage fintech.
$Semiconductors data platform — 4-layer medallion pipeline with YAML rule engine (15+ checks), watermark ingestion, and yield analytics.
$Dataset recommender across 25+ sources — discovery time 2 hrs → <5 min; CFPB analysis surfaced ~450 complaint/month trend (p<0.001).
$Crossref retraction pipeline — automated drift detection across 208 DOIs; precision/recall analysis of flag classifiers; 3 conference presentations.
Luyang Si
data-analyst && researcher
Champaign, IL
Python
SQL
Power BI
R
NLP
Azure SQL
- Data analytics & BI: End-to-end sales pipelines, KPI dashboards, and executive reporting in Python, R, and Power BI.
- Behavioral research: Decision-making under uncertainty, confidence calibration, and transcript-level semantic analysis.
- Data engineering: Medallion-layered pipelines with watermarking, incremental ingestion, and schema-drift detection.
- NLP & embeddings: Multilingual sentence embeddings, BERT, LLaMA fine-tuning for cross-language and healthcare applications.
Skills
A snapshot. View full proficiency breakdown →
Data Analytics & BI
Python
R
SQL
Power BI
Tableau
Pandas
Statistical Modeling
Data Engineering
Azure SQL
ETL Pipelines
Watermarking
Schema-drift Detection
Medallion Architecture
Git
NLP, ML & Research
Multilingual Embeddings
BERT / spaCy
LLaMA Fine-tuning
Random Forest
A/B Testing
Precision/Recall
Behavioral Research
Survey Design
Featured projects
Four fast reads for recruiters. Full list is on the Projects page.
Currently: Open to data science, data engineering, and data analyst roles. Available immediately.
Canvas Platform Data Ingestion
Medallion-layered pipeline (raw/cur/meta) for Azure SQL with watermark-based incremental loads and monitoring.
Crossref Retraction Metadata Analysis
Analyzes retraction-flag consistency and tracks indexing drift (208 DOIs) impacting research integrity.
Semiconductors Data Platform
End-to-end fab operations pipeline with 4-layer medallion architecture, YAML rule engine (15+ validation checks), watermark-based incremental ingestion, and yield/equipment health analytics.