Skip to content
View aaronfpl-dev's full-sized avatar
🎯
Focusing
🎯
Focusing
  • Tijuana, Baja California, Mexico

Block or report aaronfpl-dev

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
aaronfpl-dev/README.md

Aarón Fernández Pinto López

Data & BI Analyst · Process Automation, Tijuana, Baja California, México

I automate operational processes and turn scattered data into systems that support decisions. I am a Business Intelligence graduate from UABC, with all 350 academic credits completed and the thesis-based degree process in progress. During a six-month internship at Telnor (Telmex–TELNOR Group) I built a coordination system that reached production, three RPA robots against corporate portals, and an ETL pipeline feeding an executive dashboard.

My approach: learn the manual process first, in the field if necessary, then automate it, and document it so someone else can continue without me.


Selected work

Every repository below is a reconstruction audited for methodological soundness. While building this portfolio I re-examined my own academic notebooks and found information leakage in three of them. I rebuilt the pipelines, the metrics dropped, and I published the honest numbers. Each README documents what was wrong and what changed.

Project What it does Headline result
telecom-customer-churn Churn classification on 6,589 customers, preprocessing fitted inside each CV fold Random Forest: ROC-AUC 0.924, F1 0.735
mexico-poverty-data-warehouse Inmon-style warehouse over CONEVAL/ENIGH microdata: staging → normalized EDW → five data marts Reproduces national estimates from 309,534 records weighted to 128.9M people
iowa-liquor-sales-analysis Sales forecasting with temporal backtesting and county segmentation Seasonal baseline wins: MAPE 4.79%. Documented as a finding, not a failure
mercadolibre-pipeline End-to-end pipeline: scraping → cleaning → MySQL → interactive dashboards Selenium · pandas · SQLAlchemy · Dash
heart-disease-model-comparison Five classifiers under nested cross-validation, 920 records from four hospitals SVM 0.897 ± 0.027; uncertainty overlaps, no clear winner
fashion-mnist-gridsearch GPU hyperparameter search, 36 fits Test accuracy 0.8748
route-optimization-algorithms Dijkstra, A*, SPFA, Floyd–Warshall and BFS on one shared weighted graph Four converge at cost 5.10; BFS minimizes hops instead
bike-sharing-temporal-regression Daily demand regression with a chronological holdout and leakage-prone target components removed Ridge: MAE 875.27, R² 0.606 on the final 147 days

Stack

Languages & data: Python · SQL / MySQL · MongoDB · R
Analysis & ML: pandas · NumPy · scikit-learn · statsmodels · SHAP · nested cross-validation
Time series: temporal backtesting · seasonal baselines · academic experience with SARIMAX
Big data: PySpark · Hadoop / HDFS
Automation & RPA: Selenium · python-telegram-bot · BeautifulSoup · Google Sheets API · openpyxl
BI & visualization: Power BI · Looker Studio · Tableau · Plotly · Dash · Seaborn
Cloud: AWS (VPC, EC2, IAM) · Azure · Google Cloud
Practices: E2E testing · chaos testing · technical documentation · Git · Scrum

Certifications

Google Data Analytics Professional Certificate · AWS Academy Graduate - Cloud Foundations · EC0935 Project Work Management (CONOCER/SEP, as Project Lead) · Scrum Fundamentals Certified

English: C1 (CEFR), certified by UABC's LEL examination.


Español

Analista de Datos y BI · Automatización de Procesos, Tijuana, Baja California

Automatizo procesos operativos y convierto datos dispersos en sistemas que apoyan decisiones. Soy egresado de la Licenciatura en Inteligencia de Negocios de la UABC, con los 350 créditos completados y el proceso de titulación por tesis en curso. En seis meses de prácticas en Telnor (Grupo Telmex–TELNOR) construí un sistema de coordinación que llegó a producción, tres robots RPA contra portales corporativos y un pipeline ETL para un tablero directivo.

Mi enfoque: aprender el proceso manual primero, en campo si hace falta, después automatizarlo, y dejarlo documentado para que otro pueda continuarlo sin mí.

Proyectos destacados

Cada repositorio es una reconstrucción auditada metodológicamente. Al armar este portafolio revisé mis propios cuadernos académicos y encontré fuga de información en tres de ellos. Reconstruí los pipelines, las métricas bajaron, y publiqué las cifras honestas. Cada README documenta qué estaba mal y qué cambió.

Proyecto Qué hace Resultado principal
telecom-customer-churn Clasificación de deserción sobre 6,589 clientes, preprocesamiento dentro de cada fold Random Forest: ROC-AUC 0.924, F1 0.735
mexico-poverty-data-warehouse Bodega de datos Inmon sobre microdatos CONEVAL/ENIGH: staging → EDW normalizado → cinco data marts Reproduce los indicadores nacionales desde 309,534 registros ponderados a 128.9M de personas
iowa-liquor-sales-analysis Pronóstico de ventas con backtesting temporal y segmentación por condado Gana el baseline estacional: MAPE 4.79%. Documentado como hallazgo, no como fracaso
mercadolibre-pipeline Pipeline completo: scraping → limpieza → MySQL → dashboards interactivos Selenium · pandas · SQLAlchemy · Dash
heart-disease-model-comparison Cinco clasificadores con validación anidada, 920 registros de cuatro hospitales SVM 0.897 ± 0.027; la incertidumbre se solapa, sin ganador claro
fashion-mnist-gridsearch Búsqueda de hiperparámetros en GPU, 36 entrenamientos Accuracy de prueba 0.8748
route-optimization-algorithms Dijkstra, A*, SPFA, Floyd–Warshall y BFS sobre un mismo grafo ponderado Cuatro convergen en costo 5.10; BFS minimiza saltos
bike-sharing-temporal-regression Regresión de demanda diaria con hold-out cronológico y eliminación de componentes que revelaban el objetivo Ridge: MAE 875.27, R² 0.606 en los 147 días finales

Certificaciones

Google Data Analytics Professional Certificate · AWS Academy Graduate - Cloud Foundations · EC0935 Gestión de Trabajo por Proyectos (CONOCER/SEP, como Líder de Proyecto) · Scrum Fundamentals Certified

Inglés: C1 (MCER), acreditado por examen LEL de la UABC.


Basado en Tijuana, México. Disponible para trabajo remoto con equipos de México y Estados Unidos. aaronfernandezsec@gmail.com


Based in Tijuana, Mexico. Available for remote work with teams in Mexico and the US. aaronfernandezsec@gmail.com

Pinned Loading

  1. bike-sharing-temporal-regression bike-sharing-temporal-regression Public

    Leakage-safe temporal regression on the UCI Bike Sharing dataset

    Python

  2. heart-disease-model-comparison heart-disease-model-comparison Public

    Leakage-safe nested classifier comparison on the UCI Heart Disease dataset.

    Python

  3. iowa-liquor-sales-analysis iowa-liquor-sales-analysis Public

    Reproducible Iowa liquor sales analysis with temporal backtesting and county segmentation.

    Python

  4. mercadolibre-pipeline mercadolibre-pipeline Public

    Pipeline de datos: scraping, limpieza, MySQL y dashboards interactivos con Dash y Plotly.

    Python

  5. mexico-poverty-data-warehouse mexico-poverty-data-warehouse Public

    Survey-weighted Inmon-style warehouse for Mexico poverty indicators.

    Python

  6. telecom-customer-churn telecom-customer-churn Public

    Leakage-safe telecom churn analysis with reproducible preprocessing, model comparison and business insights.

    Python