00 — SANA VERGHESE / PORTFOLIO 2026

I turn messy
event logs into decisions
people actually make.

Senior data analyst with six years across marketplace, fintech, and healthtech teams. I build SQL models and BI dashboards that survive a schema migration, and I write the analysis that gets read in the meeting — not filed after it.

847,000

ROWS MODELED NIGHTLY IN DBT

$1.4M

ANNUAL SAVINGS SURFACED

01 — BY THE NUMBERS

23

DASHBOARDS SHIPPED TO PRODUCTION

6

YEARS IN ANALYTICS

41

A/B TESTS DESIGNED & CALLED

0.31s

MEDIAN QUERY LATENCY, EXEC LAYER

02 — SELECTED WORK

04 CASE STUDIES
marketplace retention cohort heatmap dashboard
FIG. 01 — WEEKLY RETENTION BY ACQUISITION COHORT

Retention curve rebuild for a two-sided marketplace

Replaced a single monthly-active number with a cohorted curve model; leadership cut a $310K reactivation spend that was chasing already-churned users.

  • BigQuery
  • dbt
  • Looker
financial funnel conversion bar chart dashboard
FIG. 02 — ONBOARDING FUNNEL, STEP-LEVEL DROP-OFF

Onboarding funnel instrumentation for a lending app

Re-specced 14 tracking events to survive a schema migration, then isolated a KYC step losing 22% of applicants. A copy fix recovered 6,800 completions per quarter.

  • Amplitude
  • SQL
  • Event Design
A/B experiment lift confidence interval chart
FIG. 03 — TREATMENT LIFT WITH 95% INTERVALS

Pricing experiment readout for a subscription team

Designed a Bayesian A/B test for a new annual plan, guarded against peeking, and shipped a decision doc that held: the 12% lift persisted three quarters later.

  • Bayesian A/B
  • Python
  • Statsmodels
executive KPI overview dashboard with line charts
FIG. 04 — EXECUTIVE OPERATING REVIEW, ONE SCREEN

Single-screen operating review for a healthtech exec team

Consolidated nine scattered reports into one governed Power BI model with row-level security, cutting the Monday prep ritual from four hours to zero.

  • Power BI
  • DAX
  • Data Modeling

03 — SQL & BI CAPABILITIES

SQL & warehousing

  • PostgreSQL, BigQuery, Snowflake
  • Window functions & CTEs at scale
  • dbt modeling, tests & docs
  • Query tuning & cost control

BI & visualization

  • Tableau, Power BI, Looker
  • LookML & DAX semantic layers
  • Metabase self-serve rollouts
  • Chart grammar & annotation

Analysis & method

  • Python (pandas, numpy)
  • Experiment design & A/B testing
  • Cohort & funnel analysis
  • Stakeholder decision docs

04 — CERTIFICATIONS

  • 2025dbt Analytics Engineering CertificationDBT LABS
  • 2024Power BI Data Analyst Associate (PL-300)MICROSOFT
  • 2023Tableau Desktop SpecialistSALESFORCE
  • 2022Professional Data Analytics CertificateGOOGLE

05 — WRITING

  1. MAR 26

    Retention isn't a number, it's a curve

    Why a single MAU figure hides the only signal that matters — and how to plot the curve your team will trust.

  2. JAN 26

    Instrumenting events that survive a schema change

    A checklist for naming, versioning, and testing tracking events so your dashboards don't quietly go dark.

  3. NOV 25

    Why I stopped trusting the average

    The distribution was skewed, the median told a different story, and the readout changed the roadmap.

06 — TESTIMONIALS

“Sana is the rare analyst who ends a meeting instead of extending it. She hands you the decision, not just the chart.”
PRIYA NAIR / VP PRODUCT, LEDGERLY
“Our dashboards used to lie by accident. She rebuilt the models and now the whole team trusts the numbers again.”
MARCUS OYELARAN / HEAD OF DATA, VESTA HEALTH

07 — CONTACT / FOR RECRUITERS

Hiring for a data analyst? Skim this, then send a note.

Open to senior and lead analytics roles, remote or hybrid in the EU/UK. Typical reply time is under a day. Full CV and references available on request.

LOCATION
Lisbon, PT · UTC+0
STATUS
Available Q3 2026
Dual7 template preview — this is a fictional demonstration business.
TemplateAxis One
Theme
Remix this siteExit