Email | Github | LinkedIn | JupyterHub
This server hosts live demos of my python code in Jupyter Notebooks. Click on a project below to get started!
Improving Part-of-Speech and Dependency Tagging of Pre-Modern Literary Chinese Texts. Modern NLP models are improved by hand-labeled linguistic annotations by a subject matter expert.
- spaCy for part-of-speech and named entity recognition engine;
- zh_core_web_sm for modern Chinese language spaCy models and POS tags;
- Pandas for building data structure of linguistic embeddings;
- displaCy for visualizing part-of-speech and dependency tags.
OSINT workflow for investigating eCommerce fraud networks. Maps buyers to sellers using publicly available feedback and reviews, allowing investigators to track fake review networks and money laundering paths between usernames.
- NetworkX for constructing graph network and calculating shortest distance between nodes;
- Pandas for scraping and cleaning transaction reviews;
- Matplotlib for static node and edge visualization;
- pyVis for interactive network graph visualization.
⚡️ eBay Patent Maintenance Report | Patent Classification Models ⚡️
Natural language processing and topic modeling techniques are used to cluster 1000 patent abstracts using latent dirichlet allocation ("LDA").
- Gensim for TF-IDF and Bag-of-Words/LDA models;
- TextBlob for part-of-speech tagging and noun extraction;
- pyLDAviz for interactive intertopic distance map.
⚡️StackBlitz | NFT Token Tracker | IP Asset Schema ⚡️
Decentralized application for minting intellectual property licenses on the Ethereum blockchain. Issues open source (ex. Creative Commons, MIT) or proprietary IP licenses as ERC-721 non-fungible tokens that are transferable peer-to-peer. Deployed to Ropsten test network.
- 0xCert asset management framework and smart contracts;
- Flask for serving metadata API;
- NES.css for frontend design and interface;
⚡️IP Gold Token | Patent Futures Exchange | Whitepaper | Enterprise | Deck ⚡️
Patent invalidation and prior art platform. Users earn IP Gold by submitting prior art that suggests an invention is not novel, or by wagering that a patent is valid or invalid. IP Gold creates an incentive for public participation in patent examination. Prediction markets provide a crowdsourced forecast of patent validity and a public incentive to increase patent quality by submitting invalidating prior art.