Justin Gosses

I live in Houston, Texas with my wife and dog Momo.

If you find a mistake on this website, please check out the CONTRIBUTING instructions and submit an issue or pull request to the code repository that builds it.

About Me


  • Things that could exist but don't yet
  • Natural Language Processing (treating words as data with a computer program )
  • Open source geospatial tools
  • Water based things like kayaking and paddle boarding. Also small web-based services that help me decide when and where to go.
  • Things at the intersection of geology & code
  • Sharing at Scale:
    • This includes questions like open data interface design; code citation tooling as a way to change its perceived value; group membership and behaviors in open source that impact how science gets done; different incentives between inner source and open source; platform and tooling effects on code sharing; and code dependencies as reusable metadata and a metaphor for developer skills.

Perspectives on Building Solutions With Code

A few personal opinions that define how I approach data science include:

  • I have a soft spot for things at the intersection of metadata, natural language processing, semantic tooling, and user interfaces as technology improvements have opened up a lot of possibilities there that have yet to be built.
  • I love finding ways to use data gathered for one purpose for a completely new purpose. It’s like finding free money.
  • I think data visualization is vastly undervalued due to so much of the understanding we get from it happening faster than we can cognitively grasp it.
  • I think we focus too much on applying machine-learning to things humans already do fairly often and not enough thinking about tasks machine-learning would be great at that we never do, because they would be too boring or time consuming.
  • Understanding how people interact with the analysis, tools, and products we create and how things get built or not built within the larger organizational context is more often the controlling variable on a successful product than the technical details or prediction accuracy.

Professional History

I have been a Program Manager, Data Scientist, Software Developer, and Geoscientist.

Here's a few highlights of my professional history with an emphasis on what types of code I used in each role and what types of problems were solved.


Currently, I work at Microsoft as a Senior Program Manager focused on improving the inner source experience, or use and reuse of code across organizational boundaries. I sit within Microsoft's Open Source Program Office. I'm relatively new and still gaining understanding of the role, but suspect I'll be writing Kusto (dialect of SQL), JavaScript, C#, and Python to understand current rates & patterns of code collaboration as we attempt to reduce friction and better align tooling and incentives.

NASA Contractor

I previously worked as a NASA contractor, supporting NASA’s Office of the Chief Information Officer’s Technology and Innovation division, now re-organized into IDAS in two roles.

(1.) As Principal Data Scientist data analytics team, I worked on projects that leverage modern data science, data visualization, and machine-learning approaches to solve client problems and help speed adoption of the latest tools. In the earliest form of the role, I focused on providing advise, consulting, and prototypes for internal NASA customers in finance, human resources, facilities, space technology, and other domains. Later on, I worked on enterprise level processes and tooling for data warehousing, data analytics, data catalogs, and metadata standards. I’ve done machine-learning projects in NLP (natural language processing), automated speech recognition, image recognition, as well as with tabular data in different forms. Most of my code was Python or JavaScript, but I sometimes did a little PHP, C++, and Java as needed.

(2.) I was also the technical program manager for Open-Innovation program, responsible for sharing NASA’s open data and open source code through the public facing websites data.nasa.gov, code.nasa.gov, api.nasa.gov, github.com/nasa as well as the programs and infrastructure to support internal-facing inner source activities. This role leveraged a lot of front-end JavaScript skills, python data munging, and python data engineering skills for APIs and data transformation services.


Before NASA, I worked as a geoscientist for BP and got to dig into data to answer questions that drove business questions across Canada, U.S., and Brazil. Towards the end of my time there I also got to use code to build solutions that domain specific software applications couldn't be twisted into answering via graphics user interfaces alone. I wrote data processing scripts in sed/awk when novel types of analytics were required. I also put together prototypes combining Python with ArcGIS and Excel based data inputs to identify risks to reservoir quality earlier than traditional methods.

Recent Talks

AGU 2019, Poster Presentation“Reusing Data and Metadata to Create New Metadata through Machine-learning & Other Programmatic Methods“

Remote talk SpaceApps Lilv, Ukraine 2019 – “A Tour of NASA’s Data Universe for a SpaceApps Audience“

American Association of Petroleum Geologists Annual Conference & Exhibit 2019 – “A Supervised Machine-Learning Approach to Stratigraphic Surface Picking in Well Logs from the Mannville Group of Alberta, Canada“

Houston Data Visualization Meetup 2017 – “Getting Your Data Visualizations Online“

Rice Data Science Conference 2017 – “Practical Considerations for Data Science Consulting and Innovation in a Large Organization“

Johnson Space Center Data Science Day 2017 – “The Changing Landscape of Data Visualization Tools Over the Last 40 Years“


"Image of the Orcid logo that says orcid"


Community Involvement

Houston Data Visualization Meetup – co-lead

Gulf Coast Section of SEPM (sedimentology geology) – former social media manager

"A photograph of a toy dinosaur holding Justin Gosses's business card." A helpful Parasaurolophus displays my business card.