Statistician and Data Scientist (Disparity Analytics) (Ref #: 1110413)

Estrada Consultingorporated • Sacramento, CA

1d • Remote

About The Position

Estrada Consulting, Inc. is seeking a statistician and data scientist for a public university research institute’s secure, web-based analytics platform for law enforcement traffic-stop data. The role owns the statistical methods behind the platform, translating the institute’s published methodology into tested, reproducible code and ensuring that results are clearly explained for police executives, researchers, and the public.

The platform will enable continuous statewide, agency-level, and officer-level analysis of millions of traffic-stop records reported monthly by more than one hundred police agencies. It will provide statistical tests to distinguish disparities from noise, role-based access for different user groups, and a foundation for early-intervention indicators and predictive analytics. The existing prototype uses Python, an accessible web interface, a columnar analytics engine, and Microsoft Azure hosting, and must meet WCAG 2.1 Level AA.

The engagement is estimated at about 550 hours over a nine-month implementation, roughly 15 hours per week, with the heaviest effort in months two through six. Part-time support may continue after go-live as enhancements require. Work is remote for US-based team members, with regular video meetings on Eastern time. The expected start is in the first quarter of 2027. Ongoing work over a three-year term is possible, contingent on contract award. The hourly rate is $60 to $95, depending on experience. Candidates must consent to a client-required background check. The candidate’s name and a resume summary will be included in ECI’s proposal.

Requirements

• Master’s degree or PhD in statistics, biostatistics, economics, criminology, public policy, sociology, or a related quantitative field.
• At least 6 years of applied statistical analysis, including hypothesis testing, regression modeling, and analysis of large administrative datasets containing millions of records.
• Experience measuring disparities or outcome differences between groups, with sound handling of benchmarks, effect sizes, and multiple comparisons.
• Strong Python skills, including pandas, SciPy, and statsmodels, and strong SQL skills.
• Working knowledge of R or Stata for independent verification.
• Ability to explain statistical findings and their limitations to non-technical audiences in writing and meetings.
• Reproducible research habits, including version control, documented code, and tests.
• Must be based in the United States and able to pass a background check.
• Must consent to the client-required background check, covering criminal history for the past seven years, sex offender registries, and Social Security number verification.

Nice To Haves

• Published or applied work on policing, traffic stops, racial profiling, or criminal justice data, such as Veil of Darkness, KPT hit-rate, or benchmarking studies.
• Experience with officer early-intervention systems or performance indicators for law enforcement agencies.
• Experience with DuckDB, Parquet, or other columnar analytics tools, and with disclosure control or small-cell suppression for public data.
• Experience working with a university research center or as a research partner to state or local government.

Responsibilities

• Reconcile each statistical test with the client’s published annual methodology, document differences, and obtain client sign-off on indicator rules before they are shown to an agency.
• Implement and validate analytical methods in Python, including two-proportion and chi-square tests, benchmark comparisons, Veil of Darkness tests using sunset and civil twilight, hit-rate tests on search outcomes, and anomaly detection on monthly trends.
• Design multivariable models, such as logistic regression with agency, time, and location controls, and use synthetic control or other quasi-experimental methods where specified by the client.
• Define officer-level early-intervention indicators with appropriate significance thresholds, effect sizes, minimum counts, and corrections for multiple comparisons.
• Independently recompute results in R or Stata on sample selections to verify the platform before release.
• Write plain-language methodology, definitions, and interpretation notes to accompany results.
• Work with the data engineer on data-quality rules affecting analysis, including coding of race and ethnicity, duplicates, missing times, searches, and contraband, and work with the business analyst on requirements.
• Present methods and results to client researchers and stakeholders, and support training and knowledge transfer.

What This Job Offers

• Job Type: Part-time