Gonzalo La Chica
Portfolio — Updated August 2026

Data & Analytics
Engineering.

I build data infrastructure, BI dashboards, and analytics tooling — these days as a Research Assistant at Clayton State University, alongside an M.S. in Data Science. Based in Atlanta, GA.

PythonSQLPower BIServiceNowDAXPower Query

This site is actively getting a refresh — a few sections may shift over the next few weeks as I update them.

01 — About
Gonzalo La Chica

I'm a data analyst working my way into data engineering. Right now I'm a Research Assistant at Clayton State University, building an agentic AI system that verifies extreme rainfall detection over Atlanta by cross-checking NASA ERA5 reanalysis data against NOAA ground-truth storm reports — part of my M.S. special project. It's early-stage but active: so far I've built the data ingestion pipeline (ERA5 plus climatology baselines, with config-driven regional scoping) and I'm designing an LLM-orchestrated verification agent in LangGraph that flags events using a causal drift score.

That builds on the BI and pipeline work I did previously as a HUB Analyst in the university's IT department (ETL pipelines from ServiceNow, Active Directory, and PDQ into Power BI — see the dashboards below), and on the M.S. in Data Science I'm about a year into, focused on relational design, Python, and cloud analytics. At the moment I'm heads-down studying for the Microsoft PL-300 (Power BI) exam, with DP-700 (Fabric Data Engineer) up next. Longer term, I'm aiming for an Analytics Engineer or Junior Data Engineer role, adding SQL depth, PySpark, and dbt to my Power BI foundation.

Outside of work I follow sports analytics, drink more coffee than I should, and like building things that actually ship.

02 — Selected Work
01

Bank Fraud & Scam Complaint Analytics

Python + PostgreSQL + dbt + Power BI — benchmarking Truist, JPMorgan Chase & American Express

Which bank's customers report more fraud, once you account for how big each bank is? I pulled 179,963 public CFPB complaints from 2020 to 2026, loaded them into Postgres with a Python script that checks the data before writing anything, modeled them into a star schema with dbt, and built a three-page Power BI dashboard on top.

Adjusted for deposits, Truist passed Chase in 2026: 14.8 fraud complaints per $10B of deposits vs 12.7. Every number on the dashboard is backed by 42 automated dbt tests and checked against a separate SQL answer key, which caught two real bugs before anything went live.

Architecture diagram: the CFPB complaint API feeds a Python script that validates the data and loads a raw table in Supabase Postgres in one transaction. dbt cleans it, builds a star schema with contracts and 42 tests, and Power BI reads it through a read-only user.
CFPB API → Python (validate, then load) → Supabase Postgres → dbt (star schema, 42 tests) → Power BI.
Overview page of the Power BI dashboard: header counters, KPI cards and fraud complaints per $10B of deposits for Chase and Truist.
Overview — scale, growth and the size-adjusted comparison (Power BI).

This is v2. The first version ran on Microsoft Fabric inside an organization's tenant, so nobody outside could open it. I rebuilt it on tools anyone can run, and you can clone the repo and reproduce every number.

PythonPostgreSQLdbtStar SchemaPower BIDAX
Case study → GitHub repo → Fraud rules → dbt models →
02

HUB Incidents Dashboard

Executive BI — Clayton State IT leadership

The flagship incident-management dashboard used daily by HUB leadership to monitor service performance. End-to-end build — from the data pipeline through the DAX model to the final Power BI report.

Fully operational intelligence tool with features like KPIs leadership: MTTR, SLA compliance %, open backlog, reopen rate, and a forecasted ticket-volume trendline so the team can see where the queue is headed, not just where it's been.

Medallion architecture diagram: Bronze ingestion sources flow into Silver Python and pandas ETL with anomaly detection, then into Gold layer SharePoint CSVs and Power BI dashboard. Windows Task Scheduler orchestrates the daily refresh.
Medallion architecture — Bronze ingestion, Silver refinement, Gold consumption.
IT Incident Management Dashboard in Power BI, showing total incidents, opened incidents, same-day resolution rate, average resolution days, reopen rate, SLA compliance, resolution trends, and team performance.
Executive Overview — IT Incident Management Dashboard (Power BI).
Knowledge Management Intelligence Dashboard in Power BI, showing article counts by knowledge base and department, top contributors, feedback trends, compliance overview, and a knowledge management maturity score.
Knowledge Management Intelligence Dashboard — the KM leg of the same BI suite.

Part of a multi-domain BI suite (incidents · anomaly reports · knowledge management). Names and select rows are redacted above for privacy — happy to walk through the live reports in an interview.

Power BIDAXPower QueryPythonPandasServiceNow APISharePoint
03

Appointment No-Show Risk Analytics

SQL data cleaning & Power BI dashboard — 110K-row healthcare dataset

Cleaned and modeled a 110,521-row medical appointment dataset in MySQL Workbench — type fixes, date parsing, and a window-function-based patient risk model (prior appointments, prior no-show rate) that scores every appointment into a risk tier — then built an interactive Power BI dashboard on top of it end-to-end in Power Query and DAX.

Appointment No-Show Dashboard in Power BI Desktop, showing KPI cards, no-show rate by day of week and lead time, the risk tier breakdown, top neighbourhoods, behavioral and demographic drivers, and the high-risk patient table.
Appointment No-Show Dashboard — Power BI Desktop, built without a template.
MySQL WorkbenchWindow FunctionsPower BIPower QueryDAX
Case study → Cleaning SQL → Analysis SQL →
04

ReceiptPal Database

Relational schema design & implementation

8-table normalized MySQL schema for a receipt-tracking application, including referential integrity, M:N resolution via associative entities, CHECK constraints, and automated constraint-validation scripts.

MySQLWorkbenchERDDDLNormalization
Case study →
05

Agentic Extreme-Weather Verification System

M.S. special project — Research Assistant, Clayton State University

An agentic AI system that verifies extreme rainfall detection over Atlanta by cross-checking NASA ERA5 reanalysis data against NOAA ground-truth storm reports.

Early-stage and actively in development. So far: a data ingestion pipeline pulling ERA5 and climatology baselines with config-driven regional scoping, plus the initial design of an LLM-orchestrated verification agent (LangGraph) that flags events using a causal drift score.

LangGraphLLM AgentsERA5NOAAPythonCausal Inference
03 — Education
Aug 2025 — Present

M.S. Data Science in Data Intelligence

Clayton State University

Aug 2023 — July 2025

B.S. Business Administration and Marketing

Clayton State University

04 — Certifications
  • ServiceNow CSA On hold
  • Microsoft PL-300 (Power BI) Actively studying — exam upcoming
  • Microsoft DP-700 (Fabric Data Engineer) Up next
05 — Contact

Open to Analytics Engineer, BI Developer, and Junior Data Engineer roles in the Atlanta area.

glachica@student.clayton.edu GitHub ↗ LinkedIn ↗