Engineer-turned-Data-Analyst with 5+ years of experience solving the same problem across very different environments — a regulated nuclear facility, a state government, and a commercial finance team: valuable information nobody was systematizing. That pattern is where my data practice actually started, years before it had a formal title.
I'm now focused on Data Analyst / BI Analyst roles, remote, on international teams — bringing strong business judgment (stakeholder management, identifying what's actually worth measuring) together with a technical stack I'm actively deepening: intermediate SQL (certifying), Python (Pandas, SciPy, Plotly), and Tableau/Power BI.
Every project below follows the same standard: no finding ships without being executed and verified against the underlying data. Where an earlier draft of my own analysis got a conclusion wrong — a regional finding in the Megaline project, an unvalidated revenue claim in the Showz A/B test — I left the correction in the notebook instead of quietly fixing it. That's part of the deliverable, not a flaw in it.
| Project | What it proves |
|---|---|
| Showz: A/B Testing & Hypothesis Prioritization | ICE/RICE hypothesis scoring + Mann-Whitney U validation. Confirms an 18.95% conversion lift (P=0.007) — and correctly identifies that a claimed revenue/AOV advantage (P=0.822) wasn't statistically real, a check the original analysis had skipped. |
| FoodFlow: A/A/B Funnel Validation | Bonferroni-corrected Mann-Whitney testing. Pinpointed a 38% funnel drop-off point, confirmed zero A/A control-group bias, and proved a typography change was statistically neutral — protecting the business from a false-positive UX decision. |
| Project | What it proves |
|---|---|
| Showz BI: Marketing ROI, Unit Economics & LTV | Cohort LTV/CAC/ROMI modeling across 10 acquisition channels. Found a channel absorbing 42.8% of budget while destroying capital (−$8.29/user, 38.6% ROMI) and a 5x revenue gap between Desktop and Mobile checkout. |
| Megaline: Tariff-Plan Profitability & Revenue Engineering | Rebuilt a 5-table billing engine from raw transactional data. Corrected an initial regional finding after re-testing it — NY-NJ underperforms, not outperforms, the rest of the country — and isolated a heavy-usage revenue segment within the lower-ARPU plan. |
| Instacart: Market Basket & Retention Analytics | 34.72% of platform traffic concentrated in a 2-day window; a 59.05% baseline reorder rate; and a Day 14–21 churn-risk window mapped for automated CRM retention triggers. |
| Project | What it proves |
|---|---|
| Chicago Urban Mobility: SQL + Weather-Impact Validation | SQL data extraction + Levene/t-test validation of a 21.4% weather-driven trip-duration increase (33.3→40.5 min, P≈0.0000) for a rideshare market-entry strategy. |
| Ice Online Store: Global Video Game Sales & User Score Analysis (confirm repo link before publishing) | Mann-Whitney U / t-test validation of rating differences across platforms and genres, plus regional (NA/EU/JP) demand segmentation to guide holiday inventory allocation. |
| Project | What it proves |
|---|---|
| Advanced User Data Processing & Segmentation | Customer Lifetime Value and cohort segmentation built from raw nested data structures using core Python (no Pandas) — isolates high-value under-30 users for VIP targeting. |
| DQA: Data Quality Assurance & User Profile Structuring | An 8-phase defensive ETL pipeline (string sanitization, type casting with exception handling, identity tokenization) turning malformed user records into a production-ready asset. |
| Urban Music Consumption: Springfield vs. Shelbyville | Regional listening-behavior segmentation across 61,253 streaming records, translating day/time engagement patterns into city-specific ad-timing recommendations. |
| US Used Vehicles Market: Interactive Streamlit Dashboard (confirm repo link before publishing) | A live Plotly/Streamlit app over 50,000+ listings, surfacing non-linear mileage depreciation and the 80K–150K-mile core liquidity zone for inventory valuation. |
I'm actively looking for remote Data Analyst / BI Analyst roles on international teams (US/Canada time zones welcome). Open to a conversation.
- 📧 Email: carlosacrespos@gmail.com
- 💼 LinkedIn: linkedin.com/in/carlosacrespos
⚡ Off-screen: board games, the Rubik's cube, chess, and volleyball.