Determine where to allocate marketing spend for maximum ROI across territories by analyzing revenue, costs, and marketing investment.

1. Schema Integration — Joined 6 tables (sales, products, categories, territories, campaigns) using clave_territorio and clave_producto
2. Data Cleaning — Calculated ingreso_total and costo_total per order, handled NULLs with COALESCE
3. KPI Calculation — Aggregated revenue, gross profit, margin %, and ROI % by territory
4. Validation — Reconciled totals across joins, confirmed no data anomalies

SQL (JOINs, GROUP BY, aggregations, COALESCE, NULLIF) | Relational database with 6 tables | Data validation & QA

USA leads with 75.8% ROI on $1.92M spend. Australia (49.2% ROI) punches above weight. UK underperforms with only 22.1% ROI despite $2.3M investment. All margins healthy (41–45%), but ROI divergence driven by marketing spend efficiency.

Recommendation: Reallocate ~$500K from underperforming EU/CA markets to USA/Australia for 30–40% ROI improvement.

Determine which customers drive the most value for a Latin American telecom by cleaning usage data, resolving data-quality issues, and segmenting customers by usage intensity and age.

1. Data Quality Diagnosis — Detected sentinel values (age = -999, city = '?'), impossible dates (40 records dated 2026), and confirmed duration/length nulls were Missing At Random by usage type
2. Cleaning — Replaced sentinels, standardized dates, and preserved MAR nulls as meaningful signal rather than imputing them away
3. Feature Engineering — Aggregated 40,000 usage records into a per-user profile (messages, calls, call minutes)
4. Outlier & Segmentation Analysis — Used IQR and Z-scores to identify power users, then segmented all 4,000 customers by usage (Low/Medium/High use) and age

Python (pandas, numpy) | seaborn, matplotlib | IQR & Z-score outlier detection | Rule-based segmentation | Google Colab

Medium use is the largest usage segment; a consistent minority of 21–47 users per metric are high-volume "power users" retained as an upsell target rather than cleaned away as noise. Basic plan dominates every segment, including the heaviest users, pointing to under-monetized power users.

Outliers were the opportunity, not the noise. The heaviest 21–47 users per usage metric were kept — not trimmed — since they represent ConnectaTel's clearest upsell segment.

Recommendation: Design an ultra-premium tier for these power users and target Medium use customers, already the largest segment, for migration incentives toward High use/Premium.

Determine where a development bank should invest in transport infrastructure by analyzing how urban mobility (congestion, delay) relates to economic productivity (GDP per capita, unemployment) across 15 Latin American cities.

• Data Integration — Merged TomTom traffic records with OECD city economic indicators using city and year keys
• Data Cleaning — Standardized column formats, parsed European numeric formatting, converted timestamps, filtered to 2024
• Aggregation — Grouped traffic records by city to calculate mean delay, congestion, and travel-time metrics per city-year
• Analysis — Computed a congestion-to-productivity ratio and ran correlation analysis across GDP, congestion, unemployment, and population

Python (pandas, numpy) | seaborn, matplotlib | Data wrangling & correlation analysis | Jupyter Notebook

The correlation matrix suggests that traffic jam is driven mainly by population size (r = 0.88), not GDP per capita (r = 0.28) as initially expected. Bogotá and Lima combine high traffic friction with lower economic output — the clearest case for investment. Mexico City and São Paulo show the highest absolute congestion, but it's scale-driven, not inefficiency. Montevideo and Brasília stand out as efficient benchmarks.

Recommendation: Prioritize transit investment in Bogotá and Lima for the highest expected economic return per dollar spent.

Identify corruption risk in public procurement across 260+ Mexican federal institutions by evaluating compliance with three principles: competition, transparency, and rule of law.

1. Research Support — Supported the IRC project in a research capacity, focused on interpreting procurement risk findings
2. Results Presentation — Helped translate analytical results into clear insights and presentation materials for public policy audiences
3. Stakeholder Reporting — Contributed to progress reporting to USAID as project funder

R (data analysis) | Tableau (interactive dashboard) | Public policy & governance research

Between 2018 and 2020, corruption risk increased in 147 of 247 federal institutions (59%), driven by weak competition, low transparency, and non-compliance. The tool was adopted as a reference in Mexican public policy debates on transparency and institutional integrity.

1. To analyse potential changes in the global context and aid system by 2040
2. To identify concrete pathways for organisational transformation
3. To develop tools and guidelines to support organisations in kick starting a transformative journey

Four scenarios mapped across two axes: network cooperation vs. survival of the fittest, and multipolar blocs vs. empires and conflict.

1. Trend Research — Led documentary research on global trends and drivers of change impacting the future of aid
2. . Stakeholder Facilitation — Facilitated consultations in Mexico and Colombia and coordinated multiple stakeholders (strategic partners, consulted organizations) within project governance
3. Multi-Phase Delivery — Contributed across the project's three phases: foundations (Causal Layered Analysis), scenarios, and transformation pathways

Strategic Foresight | Horizon Scanning | Causal Layered Analysis (CLA) | Scenario Building | Stakeholder Alignment | Project Management | Data analysis

Synthesizing 877+ voices from 50+ organizations — most from the Global South — into four scenario frameworks and an organizational toolkit for humanitarian resilience under high uncertainty.

Actor mapping and strategic intelligence under conditions of high uncertainty for executive decision-making.

OSINT | Strategic Intelligence | Qualitative Risk Analysis

Build a foresight-based framework to turn uncertainty into opportunity for Nuevo León's creative and cultural industries, facing accelerated digital disruption, AI integration, and post-pandemic pressure.

1. Workshop Design & Facilitation — Designed and led strategic foresight workshops with creative-sector actors and key decision-makers
2. Policy Translation — Translated the foresight exercise into an actionable public policy framework (Logical Framework Approach)

Horizon Scanning | Scenario Building | Strategic Facilitation | Logical Framework Approach

Horizon scanning report for the creative economy | 4 future scenarios for Nuevo León's creative industries | Public policy framework (Logical Framework Approach) | Published strategic policy report