Dave Gordon
Data Science in Cranford, NJ
About
Product manager with 7 years of experience diving into problems to create value. A demonstrated history of communicating between technical and non-technical stakeholders while leading complex high-pressure projects. Motivated by boiling down difficult work to a button press.
Projects
Directed the team that developed all election night data feeds that parsed and ingested election results for all 50 states to declare the next President of the United States in the 2020 General Election. Team was ultimately responsible for 85% of all new data on election night.
Produced, voiced, and edited video training materials for all vote reporters working for the NEP on election night. Presentation was created using PowerPoint animations, captured and edited using Camtasia and Audacity
Side Projects
Designed custom social media website combining some of the best features across Instagram, Reddit, and TikTok into a new twitter-style posting site. Website is intuitive to use (Instagram), has up and down votes (Reddit), and an awesome comment section (TikTok). Built using Python, Flask, SQL, AWS, JQuery, HTML and CSS.
The goal of this project was to solve the problem of fuzzy matching once and for all.
Fuzzy Matching is the processing linking similar but imperfect strings. For example "St. Louis" == "Saint Louis" and "Adams St." == "Adams Street" -- but linking datasets based on imperfect matches is time consuming to set up for one off projects.
I developed Supermatch to make this difficult process easy. To use, users paste in their two excel data sets and a csv of their matched dataset is returned. By default, the system only matches "good" matches (matches of 85% or better) -- but the user can override this if they want. There's no specific file formatting required, it doesn't matter what your headers are named, and the system can run 5 thousand matches in a second.
This project was built using Python-Flask and JQuery and is running on an AWS EC2 instance.
Work Experience
• Advance data science and automation efforts by leading development of data products, custom dashboards, and internal automation workflows.
• Analyze gigabytes of data using Python and R from MySQL databases, API endpoints, and Selenium web scraping to create insights for key decision makers and clients.
• Manage a team of engineers to scrape and ingest data feeds of election results for ABC, CBS, CNN, and NBC for the 2020, 2021 and 2022 election cycles increasing the number of election night data feeds by 462%.
• Create new diagnostic models to extract key performance indicators and cut data collection costs by 26%.
• Develop internal tools to automate pain points and optimize analysis. These tools saved 8 hours per project writing SPSS syntax, about 500 work hours per quarter tracking down information online, and converted 24-hour data runs into 2-minute scripts.
• Utilized R and MySQL to generate reports and guide both internal and client decisions.
• Trained and manage a team of nine direct reports. Delegated tasks, oversee operations and relay progress to Senior Vice President.
• Coordinated with all 50 state election administrations to ensure firm was adhering to election law with its research coverage.
• Created an application using R Shiny and Plotly to plan phone call volume cutting operator staffing by 20%.
• Designed solution to automate emailing of thousands of unique documents to team members using R, PHP and AWS S3 which removed as much as 50 hours of effort per project.
• Coordinate the Election Exit Polls for the National Election Pool in fourteen states.
• Use advanced Microsoft Excel to complete ad hoc research and data requests using pivot tables and index match.
• Hire and manage an interviewing staff of over 400 individuals.