projects

Visulization of Multilevel Modeling

A visual guide to multilevel modeling, showing how MLM handles nested data structures in cross-national research.

How Terrifying the Dream Is：A Word Embedding Approach

This project uses word embedding technology to quantitatively analyze and measure the emotional tone of dream descriptions, revealing the underlying sentiments of our subconscious narratives.

Analyzing Content Controversy in Online Communities Using Word Embedding

An experimental exploration of measuring controversy on Zhihu using word embeddings, revealing unexpected patterns in how controversial content drives user engagement.

Difference-in-Differences (DiD) in Journalism and Communication Studies

How Difference-in-Differences (DiD) analysis is applied in recent journalism and communication research

Heatmap Visualization of Intangible Cultural Heritage

This project uses Python and Baidu Map API to create a spatial heatmap visualization of China's Intangible Cultural Heritage (ICH), revealing its distribution across the country at the county level.

PopBert：Unveiling Populist Language in Political Discourse

In this project, I am learning and replicating the key steps in building the PopBERT model from the research "PopBERT. Detecting Populism and Its Host Ideologies in the German Bundestag" by Erhard et al. (2024). This study develops a transformer-based model to detect populist language in German parliamentary speeches, focusing on the moralizing references to "the virtuous people" and "the corrupt elite."

Proportional Symbol Charts in ggplot2

Using R's ggplot2, we create a Proportional Symbol Chart to illustrate ICH policy patterns across Chinese provinces over time, offering detailed control for enhanced data representation.

Sentiment Classification Using Fine-tuned BERT