# Jeová Anderson - Data, Automation & AI Portfolio

Jeová Anderson develops intelligent solutions that combine data, automation and artificial intelligence to improve processes and support decision-making. His work includes Data Analytics, Business Intelligence, Python, SQL, systems integration, RPA, Google Apps Script, Google Sheets and Gmail automation.

Website: https://geovatatsuga.github.io/portifolio/
Structured JSON: https://geovatatsuga.github.io/portifolio/portfolio.json
GitHub: https://github.com/geovatatsuga
LinkedIn: https://www.linkedin.com/in/jeova-anderson/
Medium: https://medium.com/@jeova.anderson

## Main Projects

### Automação de Rotinas Contratuais
Google Workspace automation that monitors contract deadlines and terms, sends preventive HTML alerts through Gmail and keeps an auditable history in Google Sheets. The solution uses a daily trigger and staged D-30, D-15, D-7 and D-0 alerts.
Repository: https://github.com/geovatatsuga/automacao-rotinas-contratuais-appsscript
Tags: Google Apps Script, Google Sheets, Gmail, Process Automation, Contract Management, JavaScript, Audit Logs, Preventive Alerts.

### Fraud Risk BI
Interactive BI interface for fraud-risk analysis, focused on fast reading, visual hierarchy, KPI cards and decision-ready charts for operational monitoring.
URL: https://fraud-risk-bi.vercel.app/
Tags: Business Intelligence, Data Science, Python, SQL, Dashboard, Fraud Detection.

### Detecção de Fraudes Lakehouse
End-to-end lakehouse pipeline for fraud detection, processing 6.36M+ PaySim transactions across Bronze, Silver and Gold layers. Random Forest with engineered features achieved 99.45% recall and R$ 73.21M in simulated financial savings.
URL: https://medium.com/@jeova.anderson/construindo-uma-pipeline-de-detec%C3%A7%C3%A3o-de-fraudes-com-azure-databricks-delta-lake-e-machine-learning-007af284a4d3
Tags: Azure Databricks, Delta Lake, PySpark, MLflow, Random Forest, Fraud Detection, Feature Engineering, PaySim.

### Segmentação de Clientes
Behavioral customer segmentation with RFM features, statistical transformations and K-Means. Customers were classified into 10 business subgroups and 4 clusters, delivered through an interactive Dash and Plotly dashboard.
URL: https://medium.com/@jeova.anderson/segmenta%C3%A7%C3%A3o-de-clientes-com-rfm-k-means-4e75c98f00d9
Tags: Python, K-Means, RFM, Plotly, Dash, Customer Analytics.

### Previsão de Churn
Integrated churn solution with SQL Server extraction, Python preprocessing, Random Forest training on 1,926 samples and Dash/Plotly dashboards. The model reached 81% accuracy and highlighted the main churn drivers.
URL: https://medium.com/@jeova.anderson/3-dashboards-1-objetivo-prevendo-e-evitando-churn-com-random-forest-d729f9b289a4
Tags: SQL, Pandas, Random Forest, Dash, Plotly, Churn Analytics.

### Cereus CRM
CRM-style product interface for organizing customer data, actions and business context with a cleaner workflow and analytics-ready visual structure.
URL: https://cereus-snowy.vercel.app/
Tags: CRM, Data Analytics, Business Intelligence, React, Dashboard, Customer Data.

## Generative AI & LLM Projects

### MythosEngine
Multi-agent architecture for generating long, coherent narratives with structured memory, designed to support complex scripts, books and creative production workflows.
Tags: Generative AI, LLMs, Multi-Agent Systems, Python, LangChain, Memory Systems.

### Big Bang Theory
Themed front-end project with content organization, visual rhythm and responsive interaction, designed as a polished character-driven interface that can evolve into AI character experiences.
URL: https://big-bang-theory-sigma.vercel.app/
Tags: FrontEnd Develop, React, JavaScript, Responsive UI, Character UI, AI Interface.

### Novel-to-Webtoon Pipeline
End-to-end pipeline that converts webnovel chapters into digital comics by combining LLM-based scripting with generative image workflows for visual consistency.
Tags: Generative AI, LLMs, Stable Diffusion, Prompt Engineering, Multimodal, Pipeline.