Found 7 top genes normalizing the negative values of Lyme disease data to be Absolute value of min-x over the range of max-min to keep large values higher and low values lower. We got the top genes from the duplicates among the classes of top genes with 17/18 predicted correctly in a 4 class random forest classifier model.

This report presents exploratory analysis of English text datasets used for building a next word prediction application.

PSYC-274 HW1

Exploratory data analysis of the HC Corpora dataset for the Johns Hopkins Data Science Capstone. Includes summary statistics, word frequency analysis, and plans for the n-gram prediction algorithm and Shiny app.

A five-chart interactive data story using open Australian data to explore loneliness, social connection, life satisfaction, age and gender. The visualisations show that loneliness is not only a private feeling, but part of a wider public wellbeing issue.

eveloped an executive-ready predictive modeling framework to identify and mitigate employee attrition risk using communication network integration. Leveraged behavioral network data to mathematically profile employee isolation, mapping out a critical risk threshold where disconnected employees face a 43% baseline probability of churn. Benchmarked predictive classification models to deliver data-driven risk stratification, allowing human resource leaders to deploy proactive retention interventions.

IBM 6510 Final Project