Skip to content
View CostaJr007's full-sized avatar
🎯
Focusing
🎯
Focusing

Block or report CostaJr007

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. 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
CostaJr007/README.md

Adeilson Costa

Data Engineering & Analytics | Python · SQL · ETL/ELT · Data Quality

Ottawa, ON, Canada · Open to junior data engineering and data analyst opportunities.

I build reproducible pipelines that turn APIs, public datasets, documents and time series into validated tables, analytical models and dashboards. My projects cover incremental ingestion, Bronze/Silver/Gold layers, schema validation, deduplication, data lineage and automated reporting.

B.Sc. in Applied Mathematics, with experience in Python pipelines, operations reporting and quantitative market analysis. Currently in the NPower Canada AI & Data Analytics program, working toward the IBM Data Analyst Professional Certificate.

Featured data engineering projects

A data platform combining official Brazilian power-system data with weather observations and forecasts.

  • Pipeline: incremental acquisition → Bronze raw files and manifests → typed Silver Parquet with invalid-row quarantine → Gold SQL views in DuckDB.
  • Engineering: SHA-256 content addressing, source revision history, versioned transformations, schema contracts, backfills, per-source scheduling and run-status reporting.
  • Delivery: D+1/D+2 forecasting, walk-forward validation and leakage tests, plus a Parquet star schema and a Power BI project as code (PBIP/TMDL/DAX).

Stack: Python · Polars · pandas · DuckDB · SQLite · Parquet · Power BI · pytest · GitHub Actions
Architecture · Results and limitations

Document-to-database pipelines for public US congressional trading disclosures, enriched with market data.

  • Pipeline: House/Senate filings → PDF parsing and optional OCR → shared validation and normalization → SQLite → FastAPI and analytical dashboards.
  • Engineering: incremental checkpoints with an overlap window for late filings, duplicate merging, source provenance, market-data caching and offline ingestion/API tests.
  • Delivery: searchable disclosures, sector analysis and estimated post-trade returns, with data definitions and source limitations documented.

Stack: Python · SQLAlchemy · SQLite · PDF/OCR · FastAPI · Next.js · pytest · GitHub Actions

A focused end-to-end ETL project for cryptocurrency OHLCV market data.

  • Pipeline: paginated Binance REST extraction → pandas/NumPy indicators → SQLAlchemy/SQLite tables → Streamlit/Plotly dashboard.
  • Engineering: HTTP retries with backoff, unique keys for raw-candle deduplication, scheduled execution, run history and error logging.
  • Delivery: technical indicators, alerts and Docker Compose commands for the pipeline and dashboard.

Stack: Python · pandas · NumPy · SQLAlchemy · SQLite · Streamlit · Plotly · Docker

More pipelines and applied analytics

Project Pipeline and engineering focus
Athena-SDA Public orbital histories + space-weather data → canonical time-series schema → deduplicated Parquet/CSV history → feature extraction, anomaly scoring and structured reports. Past-only model training and walk-forward validation.
Viral Formula Studio Video ingestion → captions/Whisper → transcript cleaning → deterministic ffmpeg/Python metrics → structured Pydantic/JSON profiles. Cached outputs, FastAPI, Docker image publishing with GitHub Actions and IBM Cloud Code Engine deployment.
Market News Telegram Bot RSS polling → GUID-based deduplication → keyword filtering → batched Telegram delivery. Persisted state, conditional HTTP requests, rate-limit handling, retries and rotating logs.

Engineering toolkit

  • Data processing: Python, SQL, pandas, Polars, NumPy, DuckDB, SQLite, SQLAlchemy and Parquet.
  • Pipeline reliability: incremental loads, checkpoints, schema contracts, quarantine, deduplication, lineage, retries, backfills and run logs.
  • Analytics and delivery: Power BI project files, DAX, Streamlit, Plotly, FastAPI and scikit-learn.
  • Development: Git, Docker, GitHub Actions, pytest and Linux.

Quantitative foundations

My applied mathematics and market-analysis background also informs these Python research projects:

Experience

  • Walmart Canada, Ottawa (2026–present) — overnight operations, scanner-to-database inventory reconciliation and FIFO replenishment.
  • Hidra (2024–2025) — managed Python data pipelines and reporting workflows; reduced client response times by approximately 60%.
  • Smart Prop Trader (2023–2024) — Python backtesting, risk-reporting automation and anomaly detection for derivatives.
  • JME Projetos, Brazil (2015–2019) — civil engineering internship; checklists reduced approval rework by approximately 40%, with quantities and cost estimates across 140+ sites.

Contact

Email · LinkedIn · Ottawa, ON, Canada

Pinned Loading

  1. brazil-power-netload-forecasting brazil-power-netload-forecasting Public

    End-to-end data pipeline and leakage-safe net-load forecasting for the Brazilian power system: Bronze/Silver/Gold on ONS open data, weather features, walk-forward backtests vs the grid operator, Po…

    Python

  2. congress-quant-tracker congress-quant-tracker Public

    US Congress filings pipeline: PTR ingest, post-trade returns, leaderboard

    Python

  3. crypto-etl-pipeline crypto-etl-pipeline Public

    End-to-end Data Engineering pipeline for crypto market data — ETL, SQL, Streamlit, Docker

    Python

  4. Athena-SDA Athena-SDA Public

    Satellite debris anomaly detection on TLE data, multi-detector fusion. IBM AI Builders entry.

    Python 2 2

  5. viral-formula-studio viral-formula-studio Public

    Multimodal reverse engineering of a creator's viral formula — copy, hooks and editing grammar, measured and delivered as an actionable playbook. IBM AI Builders Challenge (July 2026).

    TypeScript 2

  6. market-news-telegram-bot market-news-telegram-bot Public

    RSS market-news pipeline: polling, GUID deduplication, keyword filtering and batched Telegram delivery

    Python