Chicago, Illinois, United States
Applying graph data science, network analysis, and machine learning to identify complex financial fraud networks.
Identifying financial fraud and anomalies among companies that sell products to the U.S. government
COPA is an independent department in the City of Chicago government, separate from the Chicago Police Department. As a data scientist, I leveraged multiple government databases to identify patterns of police misconduct throughout the city. I collaborated with policy analysts to support data-driven policy recommendations to promote police accountability and public safety. • Led a three-person team audit of CPD employment databases and wrote a public report which was highlighted at City Hall and responded to by CPD’s Chief of the Bureau of Internal Affairs • Developed SQL queries to identify anomalies in COPA’s databases and shared findings with investigators with a Power BI dashboard, leading to timely and efficient closure of 10+ COPA cases • Created and deployed an interactive Power BI dashboard to guide decision making in support of investigations • Awarded Overall Excellence Award for best new hire in a non-investigative role in COPA Academy IX
Data scientist on a team analyzing satellite time series data for various U.S. government programs • Created a Python-based software package to read (pandas), visualize (matplotlib), and manipulate (pandas, NumPy, SciPy) semi-structured time series data • Developed algorithms for anomaly detection in time series data, using Python in a Linux environment • Forecast time series data using random forest regression (scikit-learn), sharing results in an internal report with government stakeholders • Shared technical findings in a variety of internal reports and presentations for stakeholders making program decisions within the U.S. government • Organized a weekly scientific seminar series for 1.5 years, attended by an average of 30 group members a week (resulted in LANL SPOT Award) • Ran and organized a highly-successful astronomy conference attended by 50 researchers (resulted in LANL SPOT Award)
Data scientist on a nine-person astronomy research project (work conducted as part of Ph.D.) • Created a Python-based software package (https://github.com/eachase/cocteau) to read (pandas), visualize (matplotlib), and manipulate (pandas, NumPy, SciPy) a dataset of over 40,000 astronomical simulations • Led a ten-person team to quantify the capability of 13 observatories to detect neutron star mergers, resulting in a first-author peer-reviewed publication with 50+ citations • Leveraged high-performance computing systems for data access and computational allocations • Developed a machine learning infrastructure based on both Random Forest classification (scikit-learn) and neural networks (TensorFlow) to infer astrophysical properties from time series data • Collaborated on four publications with observational astronomers, comparing theoretical time series data to telescope observations
For the duration of my Ph.D. studies, I worked in Northwestern's Gravitational Wave Astrophysics Group, where I used data science and Bayesian statistical analysis to uncover the properties of neutron stars and black holes throughout the universe. • Led data analysis on behalf of a 1000+ person collaboration to determine the properties of the lightest black hole merger detected to-date using Bayesian parameter estimation, with Python in a Linux environment • Announced the detection of a new binary black hole merger to the American Astronomical Society on behalf of a 1000+ person collaboration • Organized a Python-based statistical study to quantify the biases introduced by noise obscuring a time series signal, resulting in a peer-reviewed publication with 90+ citations • Mentored three high school students and one undergraduate in independent research projects • Delivered 30+ public outreach talks reaching a total audience of over 1500 people