Scientific research, analysis, and communication

Ready for your
tough data problems

I’m Dr. Sarah Sonnett, a research scientist with two decades’ experience, much of it measuring things no one can see directly: the size, shape, spin, and makeup of asteroids, worked out from faint, noisy telescope data. Nearly ten of those years were on NASA space missions, starting at NASA’s Jet Propulsion Laboratory.

I have also spent years explaining that work to every kind of audience, through public talks, workshops, and three years of teaching university courses for majors and non-majors. Through Spark Scientific, I can bring those methods to your project, whether the data is yours or comes from public sources: careful analysis, custom software, simulation, and clear communication.

Book a Free 30-Minute Consultation See the Work
Portrait of Dr. Sarah Sonnett
20+years of research experience
~10years on NASA space missions
Ph.D.in Astronomy, with degrees in Physics & Math
3+years of college teaching

Services

What Spark Scientific can do for you

These are the areas I work in most, and I’m happy to consider projects outside of them.

Signal & Image Processing

  • Finding faint signals in noisy, irregular data
  • Image processing: detection, calibration, and measurement
  • Detecting, tracking, and characterizing moving objects
  • Camera calibration and multi-view geometry
  • Removing instrument artifacts without distorting the science

Tools: Python, NumPy, SciPy, OpenCV

Examples: Spin period · Faint objects in images · 3D shape from brightness · Spectral pipeline

Modeling & Simulation

  • High-fidelity simulations of what a sensor or survey will detect, and what it will miss
  • Fitting physical and statistical models to data
  • Inferring hidden properties from indirect measurements
  • Weighing competing explanations against the evidence

Tools: Python, NumPy, SciPy, scikit-learn, ray tracing, convex shape inversion

Examples: Survey simulation · 3D shape from brightness · One rock or two?

Data Analysis & Statistics

  • Statistical analysis with honest uncertainties
  • “Did it work?” evaluations of programs and policies
  • Study design, including pre-registered tests
  • Geospatial and travel-time analysis

Tools: Python, pandas, SciPy, GeoPandas, Monte Carlo methods

Examples: Library study · Spin-period uncertainties · One rock or two?

Data Cleanup & Pipelines

  • Organizing and cleaning messy data from many sources
  • Vetting incoming data for gaps and inconsistencies, such as conflicting header fields, before any analysis
  • Automated, tested pipelines, delivered as documented GitHub repositories
  • Finding and improving code and workflow inefficiencies (in one past project, a key calculation ran about 750 times faster)

Tools: Python, git, pytest, GitHub Actions CI

Examples: Survey simulation · Library study

Communication & Training

  • Polished reports and slide decks
  • Shareable web reports and interactive demos
  • Public talks for any audience, from libraries and science museums to conference stages
  • Curriculum and workshop design, from university courses for majors and non-majors to summer camps
  • Grant proposal support

Examples: Find the Spin demo · Library access demo

Selected work

Highlights from some projects

See All Projects

Contact

Have a project in mind?

Tell me a little about your project and any data you have, or the public data you’d like examined, even if you’re not sure what’s possible yet. The first 30-minute consultation is free.

Get in Touch

See how a project runs →