Erin Weiss's Personal Portfolio

Data Scientist | AI & ML Enthusiast | Loves turning data into insight and insight into design — uniting form, function, and intelligence.

Welcome to my personal portfolio repository! This GitHub repository showcases the projects, skills, and experiences I've built as a Data Scientist. From data analysis and AI models to web development (seen through the portfolio itself) and beyond, this collection reflects my passion for creating impactful and innovative solutions.

---

Live Demo

Check out the live version of my portfolio:
👉 Portfolio URL

---

About Me

I'm a Data Scientist with a passion for AI, machine learning, and building models that make a difference.
I enjoy solving complex problems and building projects that are both analytical and creative.

In this repository, you'll find projects that highlight my skills in:

• Data analysis: Python, Pandas, SQL, R, Artificial Intelligence (AI), Reinforcement Learning, Data Visualization, Tableau, Jupyter, NumPy, Matplotlib
• DevOps and CI/CD: Git, GitHub, Bash, Shell
• Frontend development: HTML, CSS, Streamlit, some JavaScript
• Creative tools: Figma, Linearity Curve Vector Art, Sketchbook

---

Technologies

Here are some of the main technologies I've worked with:

Data Science:
Artificial Intelligence (AI), Computer Vision, Convolutional Neural Networks (CNN), Data Analytics, Data Mining, Data Modeling, Data Visualization, Deep Learning, Gurobi, Jupyter Notebook, Machine Learning, Matplotlib, Natural Language Processing (NLP), NumPy, Optimization, Pandas, PCA, Predictive Analytics, Predictive Modeling, PySpark, Python, R, Recommender Systems, Scikit-learn, Seaborn, Statistical Data Analysis, Statistical Modeling, Tableau, TensorFlow, Keras, Time Series Analysis

Databases:
Alteryx, Data Warehousing, Extract, Transform, Load (ETL), Google Analytics, Google Cloud Dataproc, Hadoop, Microsoft SQL Server, MySQL, PostgreSQL, Spark / Apache Spark, SQL, SQL Server Reporting Services (SSRS)

DevOps & CI/CD:
Bash, Git, GitHub, LaTeX, Markdown, Quarto, Shell

Frontend:
CSS, Excel (pivot tables), HTML, JavaScript, Streamlit

---

Project Highlights

Here are a few of the projects you'll find in this portfolio:

---

Reinforcement Learning Project

• Description:
  This project explores reinforcement learning through hands-on implementation of algorithms in a Jupyter notebook.
  The goal is to demonstrate how an agent can learn to make decisions over time in an environment using reward signals.
  Concepts such as exploration vs. exploitation, policy optimization, hyperparameter tuning, and reward shaping are demonstrated through code and visualizations.

• Technologies Used:
  Python, NumPy, Pandas, Matplotlib, Jupyter Notebook

• Live Demo: Demo Link

• Code: Project Folder

• Portfolio Page: Link

---

R Real Estate Project

• Description:
  An analysis of U.S. housing trends (2016–2022) using Realtor.com data.
  Explores how factors such as region, season, square footage, and market activity influence median listing prices.
  Includes rich visualizations and a multiple linear regression model explaining ~95% of price variation.

• Technologies Used:
  tidyverse, ggplot2, lubridate, summarytools, DT, ggpubr, HydroTSM, descr, SemTools

• Live Demo: Demo Link

• Code: Project Folder

• Portfolio Page: Link

---

Excel Cleaning & Comparison App

• Description:
  A Streamlit web application for cleaning, standardizing, and comparing Excel files, especially messy spreadsheets created from PDF-to-Excel invoice conversions.

  Includes a flexible General App for a wide range of invoice-style layouts and a streamlined Aftermath App tailored to Aftermath Disaster Recovery’s recurring invoice + monitoring workflows. Users can clean files, compare them using a shared ID column, detect missing or mismatched records, and download all results as multi-sheet Excel outputs. A built-in Tutorial with sample files and a Contact form support onboarding and feedback.

• Technologies Used:
  Python, Streamlit, pandas, openpyxl, Pillow, requests, email_validator, captcha, streamlit_js_eval

• Live Demo: Demo Link

• Code: Project Folder

• Portfolio Page: Link

---

Michelin Stars in Washington, D.C.

• Description:
  An interactive Tableau dashboard that explores the Michelin-starred restaurant landscape in Washington, D.C.
  Using a dataset collected in late 2023, this project compares D.C.’s fine dining scene to other global culinary hubs and visualizes trends across cuisines, restaurant locations, and Michelin ratings.

  The dashboard includes:

  • Interactive Map: Hover to view restaurant details (name, cuisine, price, website)
  • Filter Options: Filter by cuisine, price range, and number of stars
  • Custom Design: Background and UI built in Figma with glassmorphism effects

• Technologies Used:
  Tableau, Figma, HTML, CSS, Data Visualization Design, Interactive Dashboards

• Live Demo: Demo Link

• Code: Project Folder

• Portfolio Page: Link

---

📂 More Projects

Credit Analysis PCA Project

• Description:
  This project is a Principal Component Analysis (PCA) exploring financial and demographic patterns across 400 credit applicants.
  The project analyzes how variables such as Income, Credit Limit, Rating, Cards, Age, and Education relate to one another and contribute to overall variance in applicant profiles.

  By reducing the dimensionality of the dataset to key components, the analysis reveals underlying behavioral and financial drivers that distinguish applicant groups.

• Technologies Used:
  R, tidyverse, Quarto, HTML Reporting, Data Visualization, PCA

• Live Demo: Demo Link

• Code: Project Folder

• Portfolio Page: Link

---

States & Crime Clustering Project

• Description:
  This project is an unsupervised learning analysis of the USArrests dataset that uncovers hidden crime-rate patterns across all 50 U.S. states.
  The project compares K-means clustering (k = 2–5) and hierarchical clustering—with and without scaling—to evaluate how Murder, Assault, Rape, and UrbanPop contribute to natural groupings in the data.
  The analysis incorporates WCSS, the Elbow Method, and the Gap Statistic to identify an appropriate number of clusters, revealing distinct high, medium, and low-crime state profiles and showing that urbanization is not a consistent predictor of violent crime.

• Technologies Used:
  R, tidyverse, cluster, factoextra,
  K-means, Hierarchical Clustering, Scaling & Standardization,
  WCSS Analysis, Elbow Method, Gap Statistic,
  Quarto, HTML Reporting, Data Visualization

• Live Demo: Demo Link

• Code: Project Folder

• Portfolio Page: Link

---

🌍 Get in Touch

I love connecting with people who share a passion for technology and innovation!
Feel free to reach out through any of the platforms below:

• LinkedIn: LinkedIn Profile
• Email: erin.michele.weiss@gmail.com

---

Thank you for visiting my portfolio!
I hope you find something that inspires you. 😊