Project Samples

Here are some of my projects that I have designed and implemented outside of my structured coursework (student orgs, research, or for fun). There are links to the project code and work samples for each, please let me know if you would like any more information!

Forest Tree Density Mapping of Drone Images with Spatial Convolutional Neural Network

• Apr 2020 - Present
  • Standardized Aerial Forest Drone Images and extracted Density Maps resulting in input data for a Neural Network Architecture
  • Designed and implemented a Neural Network Architecture from 2017 CVPR research, utilizing perspective vectors and leave one out training data resulting in 80% prediction accuracy
  • Notable Technologies Used: Python [PyTorch, TensorFlow, NumPy], MATLAB, AWS Virtual Machine
  • Code and Writeup

Contrary presence of Competing Species Coexistence with ML Density Based Clustering Models

• Jan 2020 - Present
  • Implemented Modified Density-Based Spatial Clustering Model from 2014 research on dimensionally reduced Trait data among Competing Tree Species supporting the theory of species with similar traits coexisting despite the competitive exclusion
  • Currently developing custom R package for Density-Based Biological Clustering Analyses and Validation
  • Notable Technologies Used: R, Python, C
  • My Code

Michigan Chemical Contamination Heatmap Visualization for Core Consumer Product

• Dec 2019 - Present
  • Structured 3,000,000+ Public EPA contamination data points resulting in Concentration Data from 2012-2019 for Chemicals in Michigan Locations above Minimum Safety Levels *Built a Web dashboard of heatmap visualizations using 200K aggregated chemical contamination data points, enabling users to identify and pinpoint precise contamination values and changes across 1100+ Michigan area codes. *Notable Technologies Used: Python [Pandas, Date-Time Forecasting, PyPlot, Matplotlib], HTML, CSS, Jinja, GitHub Pages
  • My Code

Bladder Cancer Stage Classification with Machine Learning Analysis

• Jan 2017 - Jun 2019
  • Implemented and trained a series of Back Propagated Neural Networks, Linear Discriminant Analysis, Support Vector Machines, and Random Forest Algorithms – validated through ROC analysis - resulting in a model cancer classification accuracy (+80%)
  • Performed Feature Extraction and Selection on 80+ Bladder CTU Scans resulting in 7 key features used to predict cancer
  • Notable Technologies Used: C++, C, Weka, MS Excel
  • Work Samples

Country Happiness Predictive Model from Independent Data Sources

• Oct 2018 - Dec 2018
  • Structured independent climate, education, foreign policy, and economic inflation data with Time Series Analysis resulting in uniform 2016 training/ testing data for 130 countries - Data chosen to reflect non-demographic significance to happiness
  • Developed a series of Machine Learning regression and classification models to predict the happiness of a given country using variable selected training datasets of 4 key predictors resulting in a ~90% prediction rate.
  • Notable Technologies Used: R [Forecast], Python, Tableau, MS Excel [ VBA ]
  • Code and Writeup

Presence of Community Alternative Stable States from Multi-species Population/Interaction Data

• Mar 2020 - Apr 2020
  • Implemented and modulated a 3 species system with species-specific population parameters with novel interaction parameters to model empirical ecological interactions.
  • Final models demonstrated the existence of alternative stable states in the presence of a subset of ecological interactions in addition to the competition demonstrated in previous research.
  • Notable Technologies Used: MATLAB, Python
  • Code and Writeup

Systematic Bias between RateMyProfessor.Com and University Internal Evaluations

• Mar 2018 - Apr 2018
  • Scraped 72,348 data values from RateMyProfessor.com and structured data utilizing string fingerprinting to solve name disambiguation between both websites.
  • Performed multi-model regression analysis and variable means identifying the significant difference in rating distribution with skewed and flattened rating distributions within RateMyProfessor.com.
  • Notable Technologies Used: R, Python [ Beautiful Soup], ANOVA, Spyder, MS Excel