One Step Further in Visualizing Consumer Review Data
An ensemble between part-of-speech tagging and related term search algorithms.
Data science, writing, and reading notes
Xiao (Marshal) Ma
Hi there. I am Xiao (Marshal) Ma, a UC Berkeley alum and data scientist. This site collects my writing, programming projects, and reading notes.
The name ALaughingHorse comes from my Chinese name, Ma Xiao, which can be read as horse laugh or horse roar. It is a strange little translation, but it fits my personality.
Latest writing
Essays and technical notes
Dimensionality-Reduction with Latent Dirichlet Allocation
Reduce the dimension of training data by over 99.7% with less than a 3% drop in performance.
How to Generate Summaries for Review Text Without Neural Networks
Getting interpretable keywords with related terms to understand consumer reviews.
Projects
Small tools for text analysis
Text Tokenizer XM
A Python package that provides a convenient pipeline for text data preprocessing.
Context Searching
A Python package that helps you understand the context of unigram results from bag-of-words analysis.