Lisa Marie Pritchett

Recently Published

Predicting Activity Class from Wearable Sensors
Machine Learning project where I train a random forest algorithm to predict activity class using wearable sensor data. I explore ways of reducing training time and improving product design.
Exploring the Activity Classification Dataset
Exploratory analysis including Principal Components and Pairplots of the activity classification data used in my machine learning project.
Fuel Economy in Automatic and Manual Cars
A nested multiple regression analysis for effects of transmission type on fuel economy.
The Distribution of Sample Means tends to be Normal
A distribution of sample means is simulated and compared to the standard normal distribution. Quantile-quantile plots and statistical tests are used.
The Human and Financial Costs of Severe Weather in the USA
A data cleaning project of The National Oceanic and Atmospheric Administration database. They define 48 types of storms but there are more than 400 unique values for storm type in the database. I use regular expressions and logical statements to efficiently clean these data into a true factor variable with a limited number of levels. Then I create a visualization to show the number of injuries and fatalities and the financial costs of storm types in the USA.