Monterey Bay painting created using a type of artificial intelligence called a Generative Adversarial Network (GAN)

Joel Silverman

Data Scientist

Data scientist with experience in imaging, geography, and natural resources gained from working in a broad range of sectors including medical diagnostic imaging, Electric Vehicle (EV) / smartgrid technology, and natural resource management. My education includes an MS from Colorado State Univ., completion of a coding bootcamp intensive. and undergrad study in plant biology. While much of my work has been in the public sector, I'm now focused on the private sector. Some of my proudest achievements include EV charging analysis that helps shift power to renewables, graduate research for NASA using satellite images, and my work in public land managment (NPS and USFS) helping reduce harmful impacts and stretch lean budgets by applying technical analysis and automation to real-world challenges.

Projects

Photo of Powerlines

Predicting Energy Demand With Time Series Analysis

Three colleagues and I predicted next-hour energy demand using time series data. Such forecasts could help customers lower their energy bill and/or increase the proportion of their power that comes from renewable energy.

Photo to tacos

Training a Convolutional Neural Net to Select High Quality Photos for the Top of Each Restaurant's Photo Gallery

Why are the photos at the top of a Yelp gallery often blurry or dark? In this project I used simple photography principles to train a convolutional neural net to select the highest quality photos in the gallery. Using AI, restaurant owners can save time and keep their photo gallery looking its best. The same technique can add value to many crowd-sourced photo galleries.

Random Letters

Real Talk vs. Chatbots Using Natural Language Processing

How realistic are the chatbots trained on Open AI's GPT2 compared to human forum discussion? In this project I try to separate the real from the simulated using Natural Language Processing and classification models.

Map of Dalmation Toadflax Distribution in Yellowstone National Park

Master's Thesis: Using Remote-Sensing Imagery to Predict Species Distributions in Three National Parks

Created models using R and SAS software to predict species habitat within three National Parks; techniques included multi-variate regression, geospatial analysis, and accounting for spatial auto-correlation. The work examined 50+ predictor variables and interactions, including satellite-derived indices of vegetation change, fire-related change, and on-the-ground species observations. You can read the abstract or download the complete thesis from GitHub.

Skills

  • Programming & Query Languages

    • * Python
    • * R Statistical Software
    • * Bash
    • * SQL

    • Python Libraries

      • * SciKitLearn
      • * Pandas
      • * NumPy
      • * TensorFlow & Keras
      • * Matplotlib
      • * Seaborn

      • Data Science Tools

        • * Data Collection & Cleaning
        • * Feature Engineering
        • * Machine Learning Algorithms
        • * Computer Vision
        • * Neural Networks
        • * Data Visualization
        • * Regression & Classification
        • * Spatial Statistics & Geo Analysis
        • * Bayesian Modeling
        • * Natural Language Processing

        • Selected Software & Platforms

          • * ESRI ArcGIS Pro
          • * Git
          • * Tableau
          • * Microsoft Excel & Access
          • * Google Colaboratory
          • * PySpark
          • * Google BigQuery
          • * AWS EC2

Resume

EXPERIENCE

Data Scientist - Hemex Health, Portland, OR (Remote)
Nov 2022 - Present
  • Analyze healthcare data, focusing on improving accuracy and repeatability of study results.
  • Develop KPIs and analytics to assess diagnostic device performance and improve medical data integrity.
  • Collaborate broadly across the company and with partners to translate healthcare data into actionable insights for product R&D and operational improvements.
  • Use Python and R for pre-processing, feature engineering, data analysis, and visualization.
  • Presenting insights to non-technical stakeholders.
Business Analytics Intern - WeaveGrid, San Francisco, CA
June 2022 - August 2022
  • Worked as a team member of a high growth SaaS startup engaged in EV, power systems, and climate solutions.
  • Developed product strategy within the Client Solutions team by producing a fine-grained market segmentation analytics using open-source and proprietary software. This included data collection, performing complexed SQL queries, and using results to build interactive dashboards with graphs, charts, and choropleth maps. Updated 30 key performance indicator data visuals and built mapping visuals from scratch; provided filters, panning, zooming, and highlighting effects.
  • Coordinated with Software Engineering, Sales & Marketing, and other teams to meet common goals; also examined a big data EV predictive model being scaled up. Participated in Agile scrums, sprint reviews, and ticket actions.
  • Contribute to positive change in business strategy through 1-on-1 meetings, team meetings, user feedback sessions, and topic-specific meetings.
Data Scientist Fellow - General Assembly, (Remote)
November 2021 - March 2022
  • Completed 12-week, 480-hour live data science instruction, gaining proficiency in Python, SQL, and cloud ML tooling.
  • Completed 6 projects and 30+ assignments using Python programming to explore a variety of machine learning / AI techniques, algorithms, and datasets. Statistical analysis methods included regression, classification, several deep learning architectures, PCA, cluster analysis, Bayesian statistics, random forest, and time series analysis. Projects emphasized business cases and learning to collaborate in coding teams.
  • Scraped website data using APIs, cleaned data, conducted data visualization, and engineered variables. Built pipelines to automate modeling processes. Utilized Git and GitHub for version control in all work.
  • Learned data engineering principles. Practiced extract, transform, load (ETL) processes, including SQL queries of databases.

EDUCATION

Master of Science - Colorado State University, Fort Collins, CO
  • Coursework in applied statistics: Spatial Statistics & Data Modeling (machine learning), Statistics: Data Design (for both experiments and observational studies), Statistics: Data Analysis), and GIS Analysis.
  • Awarded NASA Graduate Research Assistantship building machine learning models using satellite image products to solve land management challenges for National Park Service. Conducted statistical analysis using software (R Programming Language, SAS, Excel), for study design development, prediction, and hypothesis testing. Worked with NASA scientists and NPS land managers to ensure design and results useful. Collected field data and cleaned complex datasets.
  • Thesis: developed statistical models of species distributions using remote-sensing data. Presented findings (oral and written) at conferences, symposia, and workshops. Thesis was primarily applied statistics; earned an MS in Forest Sciences through the College of Natural Resources.
Bachelor of Arts - The Evergreen State College, Olympia, WA
  • Emphasis: Natural Resource Management & Plant Biology.
  • Human & Cultural Geography, Three-month Study Abroad (Central America).

CERTIFICATES

Data Science Immersive – General Assembly, 12-weeks (full-time)
Data Analysis in SQL & Python (20+ Courses) – DataCamp, 100+ hours

SELECTED AWARDS

USFS & NPS Certificates of Appreciation
NASA Graduate Fellowship
Colorado State University Educational Event Award

PERSONAL INTERESTS

Coaching my son's soccer team.
Amateur digital photographer with a strong interest in photoprocessing using Lightroom and other tools.
Training for the Big Sur Marathon in April 2025. Completed two trail marathons (Muir Woods and Golden Gate Headlands).
Backpacking and camping with family and friends (Western US, New Zealand, Italy).
-->