Skip to content
View gikeross's full-sized avatar

Block or report gikeross

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
gikeross/README.md

Giacomo Rossini

Data Analyst | Analytics Engineering · Python · SQL · Power BI

I build analytics solutions that move from raw data to decision-ready outputs: reproducible pipelines, dimensional models, validated analysis, management KPIs, dashboards and data-driven applications.

My portfolio spans financial analytics, analytics engineering, SQL, Power BI, time-series forecasting, machine learning and OCR/NLP application development.

Featured Projects

End-to-end financial analytics engineering project for synthetic fee monitoring, reconciliation and revenue-leakage analysis. The project models the full path from transaction generation through PostgreSQL transformations and dimensional modelling to BI-ready datasets and Power BI measures.

Python · SQL · PostgreSQL · Power BI · dimensional modelling · ETL/ELT · data quality · pytest · GitHub Actions · Docker

Portfolio highlights: deterministic synthetic financial data, finance-safe fee calculations, staging/intermediate transformation layers, star-schema dimensions and facts, management KPI marts, SQL quality checks, Power BI-ready exports, DAX measure definitions and automated CI.

End-to-end OCR/NLP web application built with Flask and Google Cloud Vision. It validates image uploads, extracts handwritten text, reports OCR confidence, identifies keywords and returns related search results through a responsive frontend.

Python · Flask · Google Cloud Vision · OCR · NLP · Docker · pytest · GitHub Actions

Portfolio highlights: automated API tests, CI, secure cloud-credential handling, production Gunicorn/Docker configuration and a Render deployment blueprint.

Time-series forecasting project investigating long-term global biodiversity trends. The analysis uses walk-forward validation, compares ARIMA against a naive baseline and publishes prediction intervals rather than presenting a single deterministic long-range forecast.

Python · pandas · statsmodels · ARIMA · time series · model validation · Tableau

Portfolio highlights: selected ARIMA (4, 1, 0) achieved RMSE 0.000846 versus 0.003221 for the naive baseline — about a 73.7% improvement in the project validation — with the point forecast reaching the 0.50 threshold in 2133.

Two analytics case studies covering COVID-19 exploration and Nashville housing data cleaning. Corrected portfolio SQL is separated from the original exploratory work and paired with reproducible Python-generated summaries and charts.

SQL · MySQL · Python · pandas · CTEs · window functions · data cleaning · GitHub Actions

Portfolio highlights: joins, CTEs, window functions, views, duplicate handling, normalization and recruiter-friendly result outputs generated from the committed project data.

Sports-analytics and regression project studying how player attributes relate to overall rating across goalkeeper, defence, midfield and attack roles.

Python · pandas · scikit-learn · regression · feature engineering · pytest · GitHub Actions

Portfolio highlights: the historical notebook is complemented by a portable tested pipeline with safer parsing, position-specific regression, MAE/RMSE/R² evaluation and support for the identified FIFA 21 raw-data schema. Concrete model outputs remain intentionally dependent on restoring the source CSV rather than fabricating results.

Technical Toolkit

Analytics & BI: SQL, Power BI, Tableau, Excel, KPI definition, dimensional modelling
Data Engineering: Python, pandas, PostgreSQL, MySQL, ETL/ELT, data quality, reproducible pipelines
Modelling: scikit-learn, statsmodels, regression, ARIMA, time-series validation
Applications: Flask, REST/API workflows, Google Cloud Vision, OCR/NLP
Engineering: Git, GitHub Actions, pytest, Docker, Gunicorn

How I Approach Projects

I structure portfolio work around a business or analytical question, make transformations reproducible, define metrics explicitly, validate models against meaningful baselines, test reusable logic and expose results in a form that another analyst or stakeholder can use.

Earlier coursework and smaller exercises are preserved in archived repositories. The projects above are the work I recommend reviewing first.

Where to Start

For analytics engineering, financial analytics and BI, start with Financial Fee Analytics Platform.

For an end-to-end application, see Document Digitalization.

For statistical modelling and validation, see Biodiversity ARIMA Forecast.

For SQL and data-cleaning work, see SQL Data Analytics Portfolio.

Pinned Loading

  1. biodiversity-arima-forecast biodiversity-arima-forecast Public

    Time-series analysis and ARIMA forecasting of global biodiversity trends using Python and environmental data.

    Jupyter Notebook 1

  2. document-digitalization document-digitalization Public

    Flask application for digitizing handwritten documents with Google Cloud Vision OCR and NLP keyword extraction.

    Python 1

  3. sql-data-analytics-portfolio sql-data-analytics-portfolio Public

    Applied SQL and Python portfolio covering data exploration, cleaning, transformation and database workflows.

    Jupyter Notebook 1

  4. fifa-player-position-analysis fifa-player-position-analysis Public

    Python analysis of FIFA player attributes and positions using EDA, regression and machine-learning techniques.

    Jupyter Notebook 1