I’m a Bourdieusian data scientist. I approach data science & data analysis from both sociological and computational perspectives. In addition to applying statistical models to analyze data on social dynamics, I also do research on the reflexivity of data and society.
My Sociology dissertation explores the relationship between statistics, societies, and politics. I also have been involved in research projects on a variety of substantive topics, such as crime, deviance, and social control, trust, mobilizations, democratic politics, world society, social networks, organizations, and elites.
I also hold an MS in Computer Science and BS in Information Egineering. My computer science interests are in machine learning, algorithm analysis, and theoretical computer science. I’m particularly interested in the mathematical and philosophical fundamentals of predictive modeling.
While I apply a variety of quantitative approaches to my social science research, here I’d like to share my notes pertaining to the theoretical foundations of machine learning algorithms and statistical models. It’s my philosophy that the deeper you understand the methods you use, the better researcher you will become. Also, it is always fun for me to dig into those commonly used models!
Perceptron Learning Algorithm (with Implementation from Scratch)
Support Vector Machines (SVM) (with Implementation from Scratch)
[Neural Networks] (coming soon!)
In addition to the mathematics of data science, I’m also passionate about data wrangling and statistical analysis.
Natural Language Processing: LDA Topic Models (Python Code)
Building a Network Data Set from Scratch (Python Code)
Replication R Code of Panel Poisson Analysis (R Code)
I use a variety of software to visualize data.
Data Visualizations: A Country Network through Intergovernmental Ties
Data Visualizations: Ego Centric Networks
You are very welcome to drop me a line: hojingmao (at) gmail.com / jh2268 (at) cornell.edu