Scroll To View More
Rowan Copley, Data Science Developer in Portland, OR, United States
Rowan Copley

Data Science Developer in Portland, OR, United States

Member since August 22, 2018
Rowan has seven years of industry experience building systems that make technical, complex data easier to understand. He has built maps of the human genome and NSF-funded serious games, as well as a variety of dashboards, APIs, data pipelines, and apps. He's friendly, communicative, can iterate quickly, and is comfortable advocating for what he thinks are good choices for a project.
Rowan is now available for hire

Portfolio

Experience

  • Python, 5 years
  • Data Science, 3 years
  • Simulations, 3 years
  • React, 3 years
  • D3.js, 3 years
  • Google Maps, 2 years
  • Machine Learning, 1 year
  • Natural Language Processing (NLP), 1 year
Portland, OR, United States

Availability

Part-time

Preferred Environment

Jupyter Notebook, Pandas, Ubuntu, Git, D3.js

The most amazing...

...project I've built from scratch is an interactive map of the complete human metabolome that visualized experimental data.

Employment

  • Software Engineer

    2019 - 2019
    FinOptimal (freelance)
    • Built an OAuth2 integration module for connecting the codebase with the Quickbooks API.
    • Automated the payroll journal entry creation for several of FinOptimal's clients.
    • Worked with client to port tens of thousands of lines of code into Python 3.
    Technologies: Python, Pandas, OAuth 2, QuickBooks
  • Front-end Engineer

    2018 - 2018
    Nimbus IoT (freelance)
    • Delivered a dashboard for visualizing contaminants in water and air from a network of sensors built on Angular, Node, and D3.
    • Built a data store to centralize front-end data handling, improving code modularity and speed.
    • Worked with back-end developers to ensure that data sent to the front-end was well-structured and compact.
    • Prototyped new charts with Nimbus based on client feedback.
    • Discovered and pin-pointed the source of an in-browser memory leak using Python for data analysis.
    • Prototyped and built timeline, multi-bar, and custom pollutants chart with D3 and CSS.
    • Advised on the best way of representing data for user understanding.
    Technologies: D3.js, Angular, JavaScript
  • Software Engineering Consultant

    2018 - 2018
    Rheos Medicines (consulting)
    • Built a dashboard from the ground up for use by Rheos scientists to explore the effects of disease on the human metabolome, working with the client from design to delivery.
    • Developed the ability to switch between multiple dataset renderings, visualising the effects of different substances on the metabolome.
    • Built the front end with D3.js, Node, Express, and Google Maps API that renders in less than a second.
    • Built a pipeline with Python and Pandas which renders tiles for over 40,000 objects at seven zoom scales, taking in an Excel file and outputting image tiles for the front end.
    • Created a search and URL integration for specifying a particular metabolite and dataset.
    • Built a search functionality using jQuery to quickly find any node in the graph.
    • Modularized the rendering pipeline so that any future dataset could be easily rendered.
    Technologies: Google Maps, JavaScript, Node, D3.js
  • Software Engineer

    2017 - 2018
    NuMedii (freelance)
    • Worked on a graph-based dashboard for scientists to see results of NuMedii's proprietary enrichment algorithm and explore the relationship between drugs, genes, and diseases.
    • Co-authored a paper on drug discovery in rare and complex diseases.
    • Worked with CSO and CTO to develop the right interface for expanding disease signatures and drug enrichments.
    • Streamlined code for building and simulating the graph using D3, a 10x speed improvement.
    • Integrated MongoDB with a proxy for serving data to the front end.
    Technologies: Javascript, D3, Express, MongoDB, jQuery
  • Software Engineer

    2016 - 2017
    GoChip (contract)
    • Developed major features for React Native and React + Electron desktop apps.
    • Built a user-facing update and caching process for the desktop app.
    • Collected and analysed data with Python to help the development team make technical decisions.
    • Wrote a proxy server with Node for cryptographic keys that improved load times of our app several-fold.
    • Built and tested RESTful architecture for user administration.
    Technologies: React, React Native, Electron, JavaScript, Python
  • Data Science Mentor

    2015 - 2016
    Thinkful (freelance)
    • Helped data science students understand concepts and successfully graduate from the course.
    • Contributed updates to course materials, improving clarity.
    • Guided students through understanding of KMeans, kNN, SVM, and other ML topics.
    • Produced my own models using datasets of bike rentals, weather, GDP, and loans using the Python data science stack and MySQL.
    • Gave feedback and advice on students' keystone project, using a topic and dataset of their choice.
    Technologies: Python, Pandas, Natural Language Processing, Scikit-learn
  • Research Engineer

    2012 - 2015
    Center for Game Science, University of Washington
    • Co-developed the synthetic biology research game Nanocrafter.
    • Spearheaded a social scoring model for the game (recognized in an award).
    • Demonstrated our lab's work at DARPA Demo Days 2014 at the Pentagon.
    • Rebuilt the game physics engine to be based on Box2D.
    • Worked with Synthetic Biology researchers to create learning ramp-up for new players.
    Technologies: ActionScript, Python, Box2D

Experience

  • ReconMap (Development)

    Tests on the human metabolome create enormous amounts of data, and if you want to be able to understand and explore that data intuitively, you need to make visualizations with it. Rheos Rx approached me to design and build a version of ReconMap, which they could use to visualize information from tests they were performing.

    The metabolome was represented as a massive, static graph layout with over 20,000 nodes, and more than 20,000 edges. I did a number of tests to see how to best display this information quickly at many zoom levels and came to the conclusion that the best way was to use the Google Maps JavaScript API. Everything I built for the project centered around that: A pipeline in Python to color nodes and edges based on a dataset of experimental results, rasterize the vector image, and slice it into tiles. A JavaScript front end used the Google Maps API to serve the images, as well as search for specific reactions or metabolites.

  • Evergreens of Washington (Development)
    https://dovinmu.github.io/cascadia-maps/washington.html

    An interactive map of Washington state for exploring which trees are native to which ecoregions, and the extent of those trees' region.

  • Python Viz Notebooks (Development)
    https://github.com/dovinmu/python-viz-notebooks

    A GitHub repository for comparing Python libraries for data visualization, exploring what's possible, and discussing their strengths and weaknesses.

Skills

  • Languages

    Python, JavaScript, Java, C++, C
  • Libraries/APIs

    D3.js, React, Google Maps, Pandas, Fast.ai, WebGL, Three.js
  • Tools

    Git, Microsoft Power BI
  • Paradigms

    Data Science
  • Other

    Machine Learning, Natural Language Processing (NLP), Maps, ECharts, AmCharts, Accounting, Simulations
  • Frameworks

    Selenium, Flask, Django, React Native
  • Platforms

    Linux, Ubuntu

Education

  • Educational retreat in Neural Networks and Geographic Visualization
    2018 - 2019
    Recurse Center - Brooklyn, New York, USA
  • Bachelor's degree in Computer Science and Mathematics
    2007 - 2012
    St. Mary's College of Maryland - Maryland, USA
I really like this profile
Share it with others