• Currently working on data analytics with SQL, Python, Tableau, and Power BI
• Experience working with machine learning and neural network applications in data science
• Strengths in Python and R programming, and SQL using Postgres and SQL Server for data analytics and visualization
• Refining skills in HTML/CSS for analytical web applications
• Enjoy statistics and documentation; no one knows why.

Click here to open portfolio

email: justin.papreck@gmail.com

• I Think I've Been Poisoned... An Analysis of Water Quality in California

  • This application investigates chemical contaminants in California's drinking water, acquiring and cleaning data, then presenting findings using Tableau
  • This acquires data from csv form as well as pdf formats from different sources
  • It takes into consideration different measurement units, chemical naming conventions, and regulation differences between governmental sources
  • Presents the data regarding high contamination using Tableau to present maps and graphs, while filtering for the past 5 and 75 years

  

• Marketing_Analysis - Ad Campaign A/B Test and Analysis

  • Performs A/B Testing on an online marketing ad campaign using Z-Test to determine whether there is a significant increase in conversion rate

  

• NFL Injury Analytics

  • This application utilizes supervised and unsupervised machine learning, deep learning, and feature analysis to find patterns in NFL injuries
  • Random Forests provided a high-accuracy, high precision model on a very imbalanced dataset, whereas Neural Networks produced a high-accuracy, high-recall model
  • Additional analytics demonstrated different correlations with injuries associated to lower-body injuries and concussions

  

• Credit_Risk_Analysis: Credit Risk Analysis with Supervised Machine Learning

  • This application utilizes Ensemble Learning and Logistic Regression to predict credit risk.
  • Easy Ensemble AdaBoost outperformed the other models, including Random Forest, oversampling, and undersampling, yielding 93% accuracy

  

• Cryptocurrencies: Clustering Cryptocurrencies

  • Application of K-Means clustering to identify groups of crypotocurrencies in unsupervised learning
  • With PCA and clustering, the currencies were grouped into 4 different groups with one cryptocurrency standing out in a group of its own

  

• Movies-ETL: Extract, Transform, Load, Action!

  • This application cleans data acquired from Wikipedia and Kaggle to create a clean dataset for a Hack-a-thon

  

Fun fact since the last time I updated this:
An ingredient list for a murder-mystery dinner party?
Eggplant
Tomatoes
Potatoes
Cayenne Pepper
Goji Berries
Bell Peppers
Paprika
Tobacco
Deadly Nightshade aka Belladonna

Perhaps. They are also all members of the nightshade family, sharing the characteristic that they all contain alakaloids. Peppers, tomatoes, eggplants, and goji berries are all nightshade fruits (they developed from a flower and contain seeds), while potatoes are a nightshade vegetable. Tomatoes contain more alkaloids in the vine and leaves than the fruit itself, which is why some people have a reaction when their skin makes contact with the vines. What makes Deadly Nightshade do deadly? It has a very high concentration of these alkaloids in all parts of the plant. The other members have a much lower concentration, making them not only safe, but in some cases even helpful.