Data analyst who cleans the chaos so clinicians and operators get calm dashboards.
I chase down messy healthcare data, stress-test metrics before they ship, and build dashboards people like to use, with reliable pipelines and clear stories.
Less text, more signals: what I am learning, testing, and fixing right now.
Applying these to patient outreach and payment integrity cases.
Stabilized a churn model that was spiking on seasonal campaigns.
Missed demand swings once; now ship alerts and holiday-aware baselines with every forecast.
I care about healthcare because families make decisions with incomplete, messy data. A project with my parents' provider made me see how a clean dashboard can change a visit. I like problems where data friction hides real answers.
Built classification and forecasting models on 1M+ row healthcare ops data; validated with stratified CV (F1/AUC > 0.82); productionized simple ensembles to keep inference under 200ms.
Cut a claims pipeline from 14 min to 2 min by rewriting joins, indexing 3 tables, and batching extracts; own ELT in PostgreSQL/BigQuery feeding Tableau dashboards.
Built feature pipelines with pandas/sklearn; automated weekly reporting saving 40+ hours/month; packaged shared utils for data quality checks and drift alerts.
Built decision-focused dashboards (Tableau, Power BI) with row-level security and clear KPI definitions; replaced static reports with weekly self-serve views for ops and leadership.
Built stakeholder dashboards (Tableau/Looker) with row-level security; improved decision latency by shipping weekly refresh instead of monthly PDF decks.
Led 6–8 person pods; set QA checklists for data pipelines; ran post-mortems when a churn model underperformed and iterated to lift AUC by 6 points.
What I use to build reliable analytics: SQL-first pipelines, production-safe Python, and dashboards that are easy to maintain.
Risk scoring, cohorting, and time-series forecasting with validation that matches the business cadence.
Own ELT from source to dashboard; keep refreshes observable and recoverable.
Dashboards tuned for decisions: filters that map to workflows, annotations over animation.
Versioned notebooks, alerting on drift/outliers, and runbooks to keep stakeholders unblocked.
Core stack: SQL, Python, Tableau/Looker. Comfortable with Git-based ELT, observability, and keeping dashboards trustworthy.
Highlights include healthcare ops analytics and product-facing dashboards; numbers below are contextual, not hype.
Built demand-forecasting dashboards for clinic scheduling (Tableau + SQL); cut overstaffing hours by ~9% after 3 months. Partnered with ops to redesign metrics and sunset unused KPIs.
Built purchase-propensity model (AUC 0.84 on 250K rows, 5-fold CV) and automated weekly client reports; reduced manual prep by ~8 hours/week for the team.
Built advising and enrollment reports; mapped student support touchpoints and cut manual list prep for advisors. Measured improved follow-ups and reduced report turnaround time.
Selected projects with architecture, validation, and decision impact.
Pipeline: PostgreSQL → dbt-style ELT → sklearn ensemble → Tableau. AUC 0.84 on hold-out; reduced manual targeting time by replacing 4 spreadsheets with one dashboard.
Built an operational safety dashboard on 15 years of FAA incident data with incremental refresh and row-level filters. Cut weekly prep by ~6 hours and surfaced incident patterns a week earlier for the safety lead.
Automated daily call transcript pulls into MySQL with Selenium + Python QC checks, then pushed to a live Tableau view. Replaced copy/paste reports, saved ~25 hours/month, and reduced missed call reviews.
Built a pricing optimization model that lifted revenue by $3M annually. Tuned algorithms to surface confident price points instead of chasing noisy lifts.
Key credentials that back my core stack; depth over tool lists.
Google Career Certificates
2023
VerifiedIBM Professional Certificate
2023
ProfessionalGoogle Career Certificates
2022
ExpertMicrosoft
2022
AssociateKPMG
2023
ProfessionalProject Management Institute
2022
AdvancedIf you're working on healthcare analytics, operational forecasting, or decision support, I'd be glad to connect.
Dallas Fort Worth, TX (Remote Available)