Hi there 👋

• 🤗 Welcome to my website
• 🔭 I’m currently working on Social Network Analysis, Machine Learning, NLP
• 🌱 I’m currently learning LLM application on digital governance, reinforcement learning in autonomous system
• 👯 I’m looking to collaborate on investigative data journalism project
• 🤔 I’m looking for help with game design, VJing
• 💬 Ask me about dencentralized social graph, cybernetics, posthumanism, AI-governance and music production
• 📫 How to reach me: sz614@georgetown.edu
• 😄 Pronouns: They/them 🌈
• 🎵 Arranger (in Logic): post-punk, mathrock, shoegaze
• ⌨ Random Thoughts: in English, in Mandarin
• ⚡ Fun fact: A human really love sea 🏝 Skateboarding 🛹 Guitar 🎸 and photography 📹

As a Data Scientist 👾

• My skills: Programming & Tools: Python (Pandas, Sklearn, Matplotlib, Seaborn, NetworkX, Nature Language Toolkit), SQL (Advanced), Tableau, PowerBI, R (RCT, DID, RDD, IV), PySpark, AWS, Google Analytics, MS Office
• Specialization: supervised methods (decision tree, KNN, Ensemble methods, Naive Bayes and SVM) and unsupervised methods (PCA, Clustering, Text Mining and Analysis and Association Analysis)

As a Policy Analyst 📑

• Concentraion in comparative policy process
• Active advocator of data ethics

Project experience 💻

💡 Network Analysis of Twitter to Identify Opinion Leader, Emotion Cascade and Community Structure

• Deployed K-core decomposition to examine the community structure, applied NLP including Name Entity Recognition, Sentimental Analysis and Topic Modelling on tweets to investigate the emotion cascade

💡 How Consensus can be Involved through Innovating Voting Mechanism?

• Using PCA and UMAP to visualize the participants’ stance on a 2-dimensional map, uses Kmeans to cluster and classify group A and B, and uses centroid coords calculation to get the distance between two groups.

💡 Predicting Attitudes toward UBI in EU using Machine Learning Techniques

• Built and trained Logistic Regression, DecisionTree, SVM, RandomForest, XGBoost and GBDT to identify 5 primary indicators and 35 secondary indicators the key influential factor on the attitude of EU citizen towards UBI

💡 Using DID to Evaluate the Impact of Intensive Case Management Services

• Designed DID to construct quasi-experiment setting for causal inference to quantify the impacts

💡 Can Remittances Compensate for Parental Absence? Evaluated by the Psychological Well- being and Educational Outcome

• Developed a multivariate model in STATA incorporating Difference-in-Difference and Propensity Score Matching
• Estimated the indirect effects and direct impacts of remittances in the context of parental work-related migration on the well-being and academic achievements of left-behind children.