Giulia Governatori

𝗦𝗲𝗰𝘁𝗼𝗿 : 𝗛𝗲𝗮𝗹𝘁𝗵 (𝗣𝗿𝗲𝘃𝗲𝗻𝘁𝗶𝘃𝗲 𝗖𝗮𝗿𝗱𝗶𝗼𝗹𝗼𝗴𝘆)

𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻 :

Development of predictive models to assess cardiovascular risk using clinical data, performance optimization with overfitting management, creation of an interactive Power BI dashboard for "What-If" scenario exploration by clinicians, and translation of complex ML predictions into actionable risk categories (Low/Moderate/High).

𝗗𝗮𝘁𝗮𝘀𝗲𝘁𝘀 :

Source: UCI Heart Disease Dataset

Dataset: 303 patients, 4 international hospitals, 13 clinical variables

Target: presence (1) or absence (0) of heart disease

𝗠𝗟 𝗠𝗲𝘁𝗵𝗼𝗱𝗼𝗹𝗼𝗴𝘆 :

Environment: Google Colab, scikit-learn, pandas, numpy, matplotlib, seaborn, joblib

Train-test split: 80/20 with stratification

Data snooping avoidance: test set locked until final evaluation

5-Fold Cross-Validation on train set only for hyperparameter tuning

𝗣𝗿𝗲𝗽𝗿𝗼𝗰𝗲𝘀𝘀𝗶𝗻𝗴 𝗣𝗶𝗽𝗲𝗹𝗶𝗻𝗲 :

Numerical variables: median imputation + StandardScaler (μ=0, σ=1)

Categorical variables: One-Hot Encoding

𝗠𝗼𝗱𝗲𝗹 𝗱𝗲𝘃𝗲𝗹𝗼𝗽𝗺𝗲𝗻𝘁 :

Model 1 - Random Forest:

Initial GridSearchCV: overfitting detected (AUC gap 5.1%)

Post-optimization result: Accuracy 85%, AUC 0.946, overfitting gap reduced to 3.5%

Model 2 - Logistic Regression:

GridSearchCV with L2 regularization

Result: Accuracy 86.7%, AUC 0.951, zero overfitting

Model 3 - Stacking Ensemble:

Base models: Optimized Random Forest + Logistic Regression

Meta-model: Logistic Regression with 5-fold CV

Result: Accuracy 83.3%, AUC 0.953 (better discrimination)

𝗞𝗲𝘆 𝗖𝗼𝗺𝗽𝗲𝘁𝗲𝗻𝗰𝗶𝗲𝘀 :

Feature Importance - What-If Scenarios - Power BI Dashboard - Machine Learning - Random Forest - Logistic Regression - Stacking Ensemble

𝗗𝗲𝗹𝗶𝘃𝗲𝗿𝗮𝗯𝗹𝗲𝘀 :

- Dashboard

- PDF Presentation (FR and EN)

- Python Notebook

- Data Dictionary

- Technical Report

𝗟𝗶𝗻𝗸 :

https://tinyurl.com/4arvaaz6

𝗦𝗲𝗰𝘁𝗼𝗿 : 𝗖𝗶𝗻𝗲𝗺𝗮 & 𝗔𝘂𝗱𝗶𝗼𝘃𝗶𝘀𝘂𝗮𝗹

𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻 :

Personal survey born from 10+ years of experience as a 3D animator in Angoulême. Objective: to factually document the real working conditions of animation professionals in France, a sector often perceived as "creative and glamorous" but whose internal dynamics are rarely analyzed with data. Using data to provide factual insights useful to former colleagues, studios, schools, and public decision-makers.

𝗔𝗰𝗵𝗶𝗲𝘃𝗲𝗺𝗲𝗻𝘁𝘀 :

Anonymous questionnaire via Google Forms → aggregation in Google Sheets

Panel: 150 respondents (animators, riggers, storyboard artists, etc. from my professional network)

28 questions covering working conditions, health, salaries, mobility, industry perception

Metadata: gender, role, experience (no personal data collected)

Google Sheets cleaning & anonymization

EDA: distributions, cross-analyses

Statistical analyses: percentages, univariate, multivariate

Quick visualizations (Chart Sheets) → PDF storytelling format + dashboard

Blog post writing (problem-data-insights-call to action structure)

𝗜𝗺𝗽𝗮𝗰𝘁 :

First independent data snapshot of the hidden side of the French animation industry

Blog post generating discussions

Contribution to local debates (studios, schools, unions) on working conditions & health

Demonstration of skills in Survey Analysis, Data Storytelling, ethics & governance (strict anonymity)

Highlighting resilience: publication despite professional pressures and risks

𝗞𝗲𝘆 𝗖𝗼𝗺𝗽𝗲𝘁𝗲𝗻𝗰𝗶𝗲𝘀 :

Google Forms • Google Sheets • Survey Design • GDPR • EDA • Statistical Analysis • Data Storytelling • Research Ethics • Sensitive Communication

𝗗𝗲𝗹𝗶𝘃𝗲𝗿𝗮𝗯𝗹𝗲𝘀 :

- Dashboard

- Presentation

- Blog Post

𝗟𝗶𝗻𝗸 :

https://tinyurl.com/ftvvhux3

𝗦𝗲𝗰𝘁𝗼𝗿 : 𝗔𝗲𝗿𝗼𝘀𝗽𝗮𝗰𝗲 & 𝗔𝗲𝗿𝗼𝗻𝗮𝘂𝘁𝗶𝗰𝘀

𝗗𝗲𝘀𝗰𝗿𝗶𝗽𝘁𝗶𝗼𝗻 :

Objective: design an online portfolio reflecting a consultant's posture, technical skills, and soft skills, including business deliverables expected by Aéroworld (fictional company).

𝗔𝗰𝗵𝗶𝗲𝘃𝗲𝗺𝗲𝗻𝘁𝘀 :

Business needs analysis: formalization of the problem, stakes.

Aéroworld functional specifications (PDF): technical/functional/regulatory constraints (GDPR, Data Lake), project KPIs

Project planning: Gantt chart listing key steps (mind map validation, functional specs, mockups, technical deliverables, mockups, dashboards, articles, tutorials)

Mockups: wireframes for 3 dashboards (tech watch, Gantt, profile presentation) via Figma

Tableau / Power BI / Looker Dashboards

Training video (16 minutes) explaining the construction of a Gantt Dashboard on Power BI, Tableau, and Looker Studio in English with subtitles in 3 languages

Documentation: user guide for dashboards, GDPR & security guarantees (Opquast, W3C), tutorial in 3 languages

Portfolio integration: assembly of pages (profile, projects, deliverables, blog, projects) in WordPress with ergonomic navigation.

𝗞𝗲𝘆 𝗖𝗼𝗺𝗽𝗲𝘁𝗲𝗻𝗰𝗶𝗲𝘀 :

Project Management • Specifications • Functional Specifications • Gantt • Figma • Tableau • Power BI • Looker Studio • User Training • Technical Documentation • Web Portfolio

𝗗𝗲𝗹𝗶𝘃𝗲𝗿𝗮𝗯𝗹𝗲𝘀 :

- Online Portfolio

- Project Mind Map

- Aéroworld Specifications

- Website Wireframe & Mock-up

- Dashboard and Website Mock-ups

- Business Needs Report

- Video and PDF Tutorial in 3 Languages

- Privacy Policy

- GDPR Documentation

- Presentation

- 3 Dashboards (Profile, Tech Watch, Gantt & Quality)

𝗟𝗶𝗻𝗸 :

https://tinyurl.com/53dsf4n8