Davirson Novoa Ramírez
Finance Data Analyst
Target roles: Financial BI Analyst · Analytics Engineer
I find the variance, I trace it to root cause, I leave the control running itself, and I explain it where the decision gets made. All four, one person.
Four figures you can audit right now
Live pipeline · refreshes dailyEach one links to its proof. Click the one you believe least.
- 15+countries whose close I own
from HQ · the Americas, Europe and Asia
- 60 hrs/moof manual work eliminated
in SLB's treasury · 10 per analyst
- 2 promotionsin 26 months at SLB
from treasury to billing specialist, Argentina and Brazil
- 6projects with public evidence
code, demo or open data, one click away
Six pieces of work. Each one checkable in one click.
Pick one: it takes you straight to the evidence.
Event study over 93.4M HMDA applications: the racial denial gap, year by year, with its 95% interval.
credit-risk-mlops · ML with model governanceDEPLOYEDCredit risk that audits itself
$276.3M
in avoidable charge-offs on the test portfolio, published next to its cost
Ten gates that block any model that fails, mine included: the credit-access model shows disparate impact of 0.7639 against a 0.80 threshold, so it is not promoted.
See the projectSPY daily series, 2022–2026: none of the five strategies ends above buy & hold.
market-data-medallion · PostgreSQL + dbtIN OPERATIONMarket data platform
51
of 1,392 strategies survived the blind window
48 assets every day, and 89 quality tests before a single figure ships.
See the projectThe semantic model: one dimension, four fact tables and two aggregates that arrive summarized from dbt.
Medallion Insights · TMDL + DAXIN THE REPOPower BI report
17
DAX measures, versioned as code
A star schema you review line by line in a pull request.
See the project

The live dashboard: GDP growth for the six largest economies since 1990, the numbered shocks and the 2026–2027 forecast.
Interactive lab · 13 modelsOPEN LABLATAM macro forecasting
20
economies forecast to 2027, with their accuracy measured
Thirteen models against the naive forecast. And you can play against it too.
See the project- below the 2018 baseline · 2018 baseline · above the baseline
Financial-inclusion index by department, 2018 and 2025, on one scale: the whole country rose at once.
M.Sc. thesis · open dataTHESIS FILEDFinancial inclusion in Colombia
17 → 1
departments below the baseline, from 2018 to 2025
19 public sources, 1,123 municipalities and the result published as it came out.
See the project 



Two real screens from the demo, with data from someone who does not exist.
Next.js + Postgres with row-level securityOPEN DEMOKairo, a multi-user product
2,741
automated tests across eight CI gates
Health and money data, open in a public demo without exposing a single row.
See the project
Track record
Read the full story →See full CV →Business Consultant, FP&A · Neoris EPAM
2026 — presentI own, from HQ, the SG&A close and forecast for 15+ countries at once — the Americas and the Eastern Hemisphere. Variance against plan, against prior forecast, and in constant currency. Fully remote, from day one.
FP&A
SLB · two promotions in 26 months
2024 — 2026LATAM treasury and billing, up to leading an intern. I spotted the FX mismatch on my own initiative and modelled it in Python, and built automation that handed every analyst ~10 hours a month back — around 60 a month across the team. Revenue recognition under SOX, and internal audits on technology projects.
Finance + data
Economic research · LEE Javeriana
2023Impact evaluation on higher-education data, with international remote teams.
Data
No logo wall
Anyone can type “Python” on a CV. Here every tool links to the exact work where I used it — the repository, the report or the role. If something has nowhere to link, it is not on this list. The full list, each with its proof, is on the CV.
Stop choosing between the person who understands the business and the person who builds the data.
One hire covers the financial analysis, the pipeline and the dashboard. Finance Data Analyst, Analytics Engineer or FP&A automation. Bogotá, GMT-5, your US hours in full. I reply within one business day, in English or Spanish.
A 20-minute video call: you pick the slot in your time zone and the link arrives by email.
The fine print, in plain sight
- Built in public, verifiable figures
- This site and the projects behind it are documented as they are made, and so are the mistakes: the engineering log carries every defect found and fixed, with its root cause. And every number on this page comes from the pipeline or the public repository, and links to the artifact that proves it. See the defect log →
- Past results
- The trading research shown here is a methodology demonstration, not investment advice.
- Power BI report
- The report exists as a PBIP project in the public repository and opens for free in Power BI Desktop. There is no public embed because “Publish to web” needs a Pro license on a work tenant and makes the dataset public.
- Analytics
- The site counts page views with Vercel Analytics: no cookies, no browser fingerprint and no personal data, served from this same domain. That is why there is no consent banner to dismiss.
- Languages
- Native Spanish · English B2 · Portuguese A2.