Sonya Sawtelle
======
#### Data scientist with a strong background in math, science and programming.
###### [ [website](http://sdsawtelle.github.io) ] . [ [linkedin](www.linkedin.com/in/sonya-sawtelle-data-scientist) ] . [ [github](https://github.com/sdsawtelle) ] . [ [sonya.sawtelle@aya.yale.edu](mailto:sonya.sawtelle@aya.yale.edu) ] . [ 802 461 3429 ]
Skills
------
data analysis databases machine learning data visualization statistics technical writing
**Programming:** Python, SQL, C++, git/github, Matlab, R, command line
**Python Packages:** Jupyter, Pandas, Numpy, Matplotlib, Scikit-learn, PyMC, Scipy, TensorFlow
**Web Development:** HTML5/CSS3, d3.js, web2py, Pelican
Experience
---------
**Data Scientist, Upwork (part-time)** (2017-2018)
- Built machine learning models in Python to inform business decisions like "How long should we leave jobs open before expiring?" and "Should this job be moved to the US-only marketplace?"
- Identified two major data leaks in a pre-existing model that impacted runtime accuracy.
- Provided insights on how client and freelancer behavior evolves during the lifetime of a job post by ingesting data from PostgreSQL and performing cleaning/analysis in Python.
- Created reports/presentations to communicate insights to both technical and non-technical staff.
**Freelance Data Scientist** (2017-2019)
- Assisted a client with MCMC Bayesian parameter estimation using PyMC and corresponding visualizations for experimental chemistry data; results contributed to a [published paper](https://pubs.acs.org/doi/abs/10.1021/acscatal.9b01234).
- Enabled a client to analyze patterns in financial transactions by building a Python module to aggregate data, create summary statistics and visualizations, and automatically email reports.
- Provided a client with insights on user behavior for a nutrition/workout tracking mobile app by performing data cleaning, exploratory analysis and PCA with clustering.
**Graduate Researcher, Yale University** (2012-2018)
- Designed and executed experiments in 3 main projects, all resulting in publications.
- Performed simulation and data analysis with Python, MATLAB, and C++.
- Improved documentation and training for lab protocols, and maintained the internal lab wiki and Unix server.
- Trained and managed 4 undergraduate assistants, resulting in successful senior projects.
**Medical College Admission Test Instructor, Kaplan Test Prep** (2011-2012)
- Planned and delivered lectures on undergraduate Physics, Chemistry and Biology.
Education
---------
**Ph.D. in Applied Physics, Yale University** (2012-2018)
- Six publications (three first-author [[1](https://journals.aps.org/prb/abstract/10.1103/PhysRevB.99.054304)][[2](https://aip.scitation.org/doi/abs/10.1063/1.5051638)][[3](https://iopscience.iop.org/article/10.1088/1361-6528/aae673/meta)]), coursework in physics, engineering, and statistics.
**B.S. in Physics, Indiana University** (2008-2011)
- Baccalaureate with Departmental Honors and Highest Distinction, 3.98/4.0 GPA
**MOOCs**
- Deep Learning 5-Course Specialization (deeplearning.ai), [Credential ID JKFBV29P5MPS](https://www.coursera.org/account/accomplishments/specialization/JKFBV29P5MPS)
- Machine Learning (Stanford University)
Awards
------
- **Sterling Prize Fellowship**, Yale University (2013). Awarded to 30 out of 10,500 applicants.
- **IU Founders Scholar**, Indiana University (2012)
- **Baccalaureate with Highest Distinction**, Indiana University (2012). Granted to 5 students.
Projects
--------
**Web App** [*GeekBuddy Social Graph and Buddy Finder*](http://sdsawtelle.pythonanywhere.com) (2017)
- Designed and built a web application using Web2Py framework and d3.js that displays the social graph for a user on www.boardgamegeek.com and colors nodes by correlation to the user as computed from boardgame ratings. Data was collected by a combination of web scraping and API querying and stored as a sqlite database.
**Website & Technical Blog** [*Exploring Data Science and Python*](http://sdsawtelle.github.io/blog/output/index.html) (2015-present)
- Designed and built a personal website for hosting useful technical articles using HTML/CSS and GitHub Pages. Authored a series of educational posts on various programming and data science topics using Jupyter Notebooks and Pelican page generator.