From complex data to useful decisions.

Four case studies showing my work in business analysis, BI, AI product quality and customer insights — without exposing confidential data.

Projects connecting data, business and product

Anonymised case studies focused on method, decisions and outcomes.

Business decisions

Connect indicators to a business question and a concrete action.

Data quality

Validate data and make indicators reliable enough to support decisions.

AI product

Test AI workflows, qualify incidents and make production releases safer.

Customer insights

Turn market data and customer feedback into actionable recommendations.

AI product

Making an AI product more reliable before production

LLM QAAPISalesforceLinear
!
Challenge
Secure AI-product workflows where edge cases, integrations and LLM outputs could degrade the experience before production.
↗
Approach
Turn test feedback into actionable tickets and create a common language across product, technical and operations teams.
✦
Contribution
I tested workflows in the Sandbox, checked API and Salesforce integrations, evaluated LLM prompts and followed 200+ incidents in Linear through to resolution.
✓
Outcome
Continuous visibility on workflow quality and more usable feedback to prioritise fixes before deployments.

Key figures

Non-confidential information
200+
Incidents tracked
Dev → Prod
Stages
Sandbox & API
Scope
My approach⌄
  • Formalised test scenarios covering business workflows and edge cases.
  • Checked LLM outputs for consistency and usability.
  • Qualified, documented and tracked incidents in Linear.
  • Coordinated feedback across Sandbox, API, Salesforce and internal teams.
View case study→
EDHEC Business SchoolCustomer insights

Identifying the signals that make a customer review helpful

PythonXGBoostNLPCustomer insights
!
Challenge
Understand which parts of a product review genuinely help other customers make a decision, beyond the star rating alone.
↗
Approach
Build a feature set combining transactional information and text characteristics, then compare their predictive contribution.
✦
Contribution
For my thesis, I analysed 50,000 Amazon baby-product reviews, created 17 variables and trained an XGBoost model to estimate perceived review helpfulness.
✓
Outcome
Verified purchases and images emerged as the strongest signals, followed by review length.

Key figures

Non-confidential information
50,000
Reviews analysed
17
Variables
XGBoost
Model
My approach⌄
  • Prepared and explored a sample of 50,000 product reviews.
  • Created variables connected to content, credibility and publication context.
  • Trained and interpreted an XGBoost model.
  • Translated findings into implications for marketing and e-commerce teams.
View case study→
Data quality

Making retail and e-commerce KPIs meaningful

Power BIRetailE-commerceData governance
!
Challenge
Make retail and e-commerce indicators comparable and useful for teams with different operational needs while supporting data quality.
↗
Approach
Combine KPI monitoring with close coordination between business and IT stakeholders to clarify definitions, needs and priorities.
✦
Contribution
I contributed to tracking indicators for four houses in Power BI, data-governance work and coordination between business and technical teams.
✓
Outcome
A more shared view of KPIs and stronger support for retail and e-commerce steering routines.

Key figures

Non-confidential information
4
Houses
Retail + e-com
Channels
Power BI
BI tool
My approach⌄
  • Tracked and consolidated retail and e-commerce indicators in Power BI.
  • Contributed to data-governance and data-quality topics.
  • Coordinated business needs, reporting and IT teams.
  • Adapted analysis to the contexts of four houses.
View case study→
Business decisions

From dashboards to category decisions

Power BINielsenKantarCategory management
!
Challenge
Help a category team distinguish genuinely useful signals across multiple sales and consumer-data sources, including in the analysis of Carrefour’s exit.
↗
Approach
Connect market and commercial-performance KPIs to a clear business question, then make the analysis accessible through reusable dashboards.
✦
Contribution
I produced more than 50 dashboards for two brands, monitored 10+ sales and consumer KPIs through Nielsen and Kantar, and helped break down the impact of Carrefour’s exit.
✓
Outcome
More structured analysis for category discussions and a shared basis for interpreting performance changes.

Key figures

Non-confidential information
50+
Dashboards
10+
KPIs tracked
2
Brands
My approach⌄
  • Built and updated dashboards from sales and consumer indicators.
  • Analysed market, category and brand-performance trends.
  • Read the impact of a distribution change through a decomposition approach.
  • Presented results in a format suited to the needs of a business team.
View case study→