Using OpenStreetMap data from the osmdata package in R, ggplot2, and other tidyverse packages I create a basic map that could be wall art. By extracting points, lines, and polygons for roads, buildings, and water features within a selected area, it is easy to see how this application could extend beyond wall decor to practical insights across industries. In future projects, I intend to explore the application of OpenStreetMap to the construction of raster files which could feasibly be used as an input feature or explanatory variable for computer vision or other spatial machine learning models

The project within this repo is from a group project, completed as a part of a Master of Science in Data Science. The project used R and .Rmd and multiple key data science packages useful for analysis, visualization, wrangling, modeling, and statistical analysis. The project insights are communicated through a report and a brief presentation. A report was also generated using R Markdown. We analyze a dataset “Default of Credit Card Clients Dataset”which contains data related to credit card statements from clients in Taiwan during certain months in 2005.Credit card defaults can have a significant impact on a lender's financial performance. When a credit card borrower defaults on their payments, the lender may not be able to recover the full amount owed, which can result in a loss. By predicting credit card defaults, lenders can take proactive measures to minimize their risk and potentially avoid losses. We predicted credit card default through the use of through meta-analysis of multiple logistic regression models combined various modeling tuning techniques, resulting in a model selection and recommendation contingent upon credit card company error costs or intervention costs.

Publicly available information is compiled and merged with geo-spatial data. Using leaflet a tool was created to help visualize the geo-spatial job data. Filtering and visualizing publicly available job data from a single company can help identify open jobs of interest at a single company. Alternatively this could be used to support informing a potentially uncertain relocation decision.

This was a collaborative project with a group of peers from the Master's in Data Science coursework. Using a subset of the well known diamond data set, the team worked together to assemble a report with relevant data visualizations to explain data nuances and develop a simple linear diamond pricing model.

There are plenty of used car data projects out there, but used car data is much more dynamic than a static data set. This project sought to be able to collect, clean, and collate publicly available used car data. Executing this project once can create a snapshot for follow-on distribution. Iterative executing used car data collection can support understanding conditions and trends within the used car world. The data could also support follow on machine learning projects.

Publicly available information from a Real Estate Website was sourced, transformed, filtered, and joined with zip-code shape-file data to create an interactive chloropleth map of five year home value index growth rates.