harveylao Goto Github PK

evidationdatasciencemodule

scrapy-tutorial-starter

starter project for my scrapy tutorial

spring-2018

PSTAT 134/234: Statistical Data Science

time-series-of-la-ozone-concentration

Watching the ozone layer in our society is an important task because the ozone layer is what protects the earth from dangerous ultra-violet rays. With the recent Thomas fire so close to SB and five other fire surrounding the LA area, ozone concentration is an ironically and incredibly relevant trend to study. This report is an analysis of ozone concentration in Los Angeles between 1955 and 1972. The goal of this study is to build a model that can accurately forecast parts per million of ozone in LA. By forecasting ozone concentration, we can better prepare for the future and investigate the variation and direction of ozone in Los Angeles. Using this time series data, we can apply statistical methods to measure trend and seasonality. We can fit autoregressive models, analyze seasonality, and apply transformations and differencing. Other methods in this study include testing residuals models for normality, variance reduction, and testing causality and invertibility. After obtaining an appropriate model, a whole year of predictions are forecasted and tested on the last 12 sample observations.

topic-modeling-genre-by-lyrics

The purpose of this project was to use R to explore the relationship between genre and lyrics using a public dataset with over 300,000 songs. The goal of the project was to discover if there was any correlation between lyrics and genres. The process included using different classification models and comparing accuracies. Methods I used include topic modeling, term frequency, LDA, GLM, stemming, and more.