Emma Ozanich, PhD
Industry and data scientist with 8 years development and 3 years leadership experience using statistics and machine learning in Python and R.
Education
Ph.D. Acoustical Oceanography | University of California, San Diego | 2020 |
M.S. Acoustical Oceanography | University of California, San Diego | 2017 |
B.S. Physics | Hamline University, St. Paul, Minnesota | 2014 |
Experience
JASCO Applied Sciences | Project Scientist, Remote (Denver, CO) | 2022 - Present |
- Lead data analysis development of predictive models and visualizations for environmental data, using Sourcetree/Bitbucket version control
- Merge and summarize data tables to visualize key outcomes with R data table, ggplot
- Model linear, nonlinear regression and statistical percentiles for timeseries data
- Communicate outcomes and collaborate with external stakeholders to align data science objectives with business goals
- Serve as technical expert in meetings with cross functional industry and government stakeholders
- Author technical memos and reports for publication in public documents
- Create software routines to improve data interpretation and efficiency
- Developed data reporting tools for accelerating large report generation
- Created interactive Python Panel dashboard for visualizing datasets with seaborn, matplotlib
Woods Hole Oceanographic Institution | Postdoctoral Investigator, Woods Hole, MA | 2021 - 2022 |
- Modeled and analyzed ocean timeseries data in MATLAB
- Used large-scale, GPU-accelerated predictive modeling to investigate environmental questions
- Cleaned and validated data to remove outliers, sub-sampled and filtered raw timeseries to select signals of interest
University of California San Diego | Graduate Researcher, San Diego, CA | 2016 - 2020 |
- Designed machine learning models for acoustic sensing in Python Keras and Tensorflow to improve predictive accuracy
- Used deep learning to estimate source angle with over 90% accuracy, robust to random noise
- Developed a deep clustering model for coral reef sound datasets using convolutional auto-encoder in Keras and Scikit-Learn
- Presented research results at semi-annual conferences and published in reputable acoustics journal
Graduate T.A. | Electrical and Computer Engineering (ECE 228) | 2019, 2020 |
Research Mentor | Haliçioglu Data Science Institute | 2018 |
- Created and presented Jupyter Notebook demo analyzing BiqQuery timeseries data with SARIMA
Skills
Languages: R, Python/Jupyter Notebook, IDL, pSQL, linux/unix, MATLAB
Development: Sourcetree, Bitbucket, git CLI, familiar with Databricks
Reporting: Microsoft Word and Excel (VBA), LaTex, ArcGIS/QGIS