Roger González March

Data Scientist | Research Engineer | Complex Science Researcher

Barcelona, Catalonia

About

Applied ML researcher and physicist focused on the intersection of artificial intelligence, data systems, and computational creativity. At the Barcelona Supercomputing Center (BSC), I contribute to Horizon Europe projects such as vCity and RELISH, developing AI-driven tools that connect urban systems, cultural data, and scientific research. My work combines state-of-the-art machine learning architectures, including large language models and complex networks, with robust data infrastructures such as PostgreSQL/PostGIS, DuckDB, and Neo4j to build full analytical pipelines. I specialize in deploying research prototypes from data to frontend applications, enabling interpretable, scalable, and visually engaging AI systems for science, culture, and society.

Experience

Barcelona Supercomputing Center

Research Engineer | Data Analytics & Visualization Group

2024 - Actual

Conducting interdisciplinary AI research within the Data Analytics & Visualization Group at the Barcelona Supercomputing Center. I contribute to two major Horizon Europe projects: vCity, where I develop a vertical line for the digital twin infrastructure to model and simulate urban systems; and Relish, where I design computational tools to study the phylogenetic evolution of recipes and cultural food heritage. I integrate advanced AI models, from LLMs to complex network algorithms, with database systems such as PostgreSQL/PostGIS, DuckDB, and Neo4j to create complete research pipelines. My work bridges backend computation and frontend visualization, ensuring that scientific insights are accessible, interpretable, and actionable.
Eurecat

Internship Data Scientist

2024 - 2024

Engaging in the 'Complexity of Information Dynamics' project to analyze disinformation spread on social networks. Reviewing AI/Deep Learning models for disinformation detection, developing deep learning models to predict content virality, and visualizing network data for trend analysis.
Universitat Pompeu Fabra

ML Researcher for Planetary Health

2023 - 2024

Applying machine learning and NLP techniques for analyzing academic and research agendas. Conducted network analysis to map the relationships between Sustainable Development Goals and UPF's academic outputs. Implemented BERT-based language models and employed data mining to uncover complex data patterns.
SDG Group

Data Analyst - Tableau Developer

2022 - 2023

Specialized in data visualization and analytics in one of the largest pharmaceutical companies, turning complex data into actionable insights for business and technical audiences. Leveraged Tableau, SQL, and Python to simplify data access and improve integration. Created impactful reports and dashboards, resolved technical issues, and collaborated with international teams using Agile methodologies.
Universitat Politècnica de Catalunya

Data Analyst at CERN n_TOF Facility

2021 - 2022

Developed and implemented digital signal processing algorithms using the C++ ROOT framework. Implemented digital filters for neutron detectors, reconstructed spectra under high-pileup conditions, and analyzed measurement data to evaluate electronics behavior.
CCiTUB

Raman Spectroscopy Researcher

2020 - 2022

Conducted research using Raman spectroscopy techniques, developing skills in scientific analysis and instrumentation.

Awards

First Prize at the Madrid Open Data Reuse Contest 2025
Ayuntamiento de Madrid

2025

Developed 'Rampa: Rutas para personas con movilidad reducida', an interactive web application that visualizes and optimizes accessible routes across Madrid. The project leverages open data and computational methods to promote urban inclusivity and improve mobility for people with reduced mobility. Won the First Prize in the Visualization category of the Madrid Open Data Reuse Contest.
Hackathon Winner at the Open Data Bizkaia AI Datathon 2024
Lantik S.A., BizkaiaTech, Bizkaiko Foru Aldundia / Diputación Foral de Bizkaia

2024

Developed an application analyzing 20,000 reviews to create an interactive map of Bizkaia's beaches, ranking them based on visitor sentiments. The project combined open data, artificial intelligence, and innovative perspectives on the region to promote tourism and gain deeper insights into the visitor experience.
Finalist at the Sahlgrenska Global Health Hackathon 2025
Sahlgrenska Global Health Hackathon, Barcelona

2025

Developed 'SmartCAP', a real-time monitoring system to track healthcare facility saturation and detect potential local outbreaks. The project aimed to create a digital twin of Barcelona’s urgency centers to predict peak times, optimize resource allocation, and improve accessibility, using technologies such as Next.js, React, Mapbox API, OpenRouteService API, and R Shiny.

Education

Universitat Pompeu Fabra, Barcelona

2025 - 2028

PhD in Information and Communication Technologies
Universitat Pompeu Fabra, Barcelona

2023 - 2024

Master's Degree in Intelligent Interactive Systems
Universitat de Barcelona

2017 - 2021

Bachelor's Degree in Physics

Projects

Rampa - Accessible Routes for Reduced Mobility •

Developed Rampa, an interactive web application that identifies and visualizes accessible routes in Madrid for people with reduced mobility. The platform combines open data from the Ayuntamiento de Madrid with computational methods to enhance urban inclusivity and mobility equity. Proposed for the First Prize in the Visualization category of the Madrid Open Data Reuse Contest 2025.

Accessible Route MappingOpen Data IntegrationUrban Inclusivity and Mobility JusticeWeb Visualization with Mapbox and Next.js
Expanded Voices - AI Voice Synthesis and Identity Exploration •

Expanded Voices is an artistic-scientific installation developed at the Barcelona Supercomputing Center (BSC) in collaboration with artist Maria Arnal. The project explores AI voice synthesis and its implications on identity, visualizing the latent spaces of timbre transfer models to reveal how artificial intelligence reshapes our sense of self. As part of the Data Analytics & Visualization Group, I contributed to the real-time visuals for Maria Arnal’s live show using TouchDesigner, implementing object and depth tracking with YOLO and Depth Anything on an NVIDIA Jetson module. Presented at Sónar+D 2025 and Ars Electronica 2025, the project merges science, technology, and art through immersive, interactive experiences.

Interactive AI Voice Synthesis InstallationVisualization of Latent Spaces in Unreal Engine 5Real-Time Visuals Developed with TouchDesignerYOLO and Depth Anything Integration on NVIDIA JetsonS+T+ARTS Residency Collaboration with Artist Maria ArnalExhibited at Sónar+D 2025 (Barcelona) and Ars Electronica 2025 (Linz)
SmartCAP - Sahlgrenska Global Health Hackathon Finalist •

Developed SmartCAP, a real-time healthcare monitoring system to track healthcare facility saturation and detect potential local outbreaks before they escalate. The project was built as part of the Sahlgrenska Global Health Hackathon and aims to predict peak times, optimize resource allocation, and improve accessibility for elderly patients using real-time data and innovative tools.

Real-Time Healthcare MonitoringDigital Twin of Healthcare FacilitiesMapbox API for Real-Time VisualizationOpenRouteService API for Accessibility AnalysisR Shiny for Outbreaks DashboardHackathon Finalist
Ranking de Playas de Bizkaia - Hackathon •

A personal project developed end-to-end, featuring data scraping, sentiment analysis, and modeling based on Google Maps reviews. The project includes an interactive Mapbox map and a scrollytelling interface to present the results intuitively and engagingly.

hackathon WinnerData ScrapingSentiment AnalysisInteractive MappingScrollytelling
Traffic Cam Photobooth BCN •

A web app that helps users find the nearest traffic camera in Catalonia (especially Barcelona) and take a selfie with it! 🎥📸 This project aims to raise awareness about constant surveillance in urban areas in a fun and engaging way.

Urban SurveillanceWeb ApplicationGeolocationPublic Awareness
Climate Shelters in Barcelona •

Storytelling with reveal.js and CARTO on the accessibility of climate shelters in Barcelona, focusing on vulnerable populations during heat waves. The research assesses current coverage, identifies gaps, and proposes solutions to improve access to shelters by 2030.

Urban PlanningClimate ResilienceSocio-demographic AnalysisCARTO
Twitter Rumor Detection with BERT Adapter •

Research project using the BERTweet model and custom adapters to classify the stance of tweets in relation to rumors.

NLPBERTTransfer LearningSupervised ML
Evaluating ChatGPT for Improving Scientific Writing Based on Readability

Research paper published in MERE 2023, examining the use of ChatGPT to enhance scientific readability through various prompt engineering and statistical techniques such as zero-shot, few shot, Shapiro-Wilk test and Mann-Whitney test.

Prompt engineering techniques Readability metricsCoherence metrics
Find My Kebab •

An interactive map that classifies kebab shops in Barcelona into color-coded tiers based on reviews and sentiment analysis.

Web scrapingGoogle Maps APIMapboxOpenStreetMapSentiment analysisTopic modeling
Data Analysis of 3He Gas-Filled Detectors

A research paper detailing a digital pulse analysis methodology for 3He gas-filled neutron counters at the CERN n_TOF facility. The study utilized fast digitizers, HTCondor for high-throughput computing, and signal differentiation techniques for improved event identification and calibration.

Digital pulse analysisCERN n_TOFHTCondor/SlurmC++ROOTSignal differentiation

Skills

Python
Jupyter
Pandas
Scikitlearn
PyTorch
TensorFlow
spaCy
SQL
AWS
Snowflake
Tableau
Git
SLURM
HuggingFace
HTML
Mapbox
CARTO
GeoPandas
Networkx
osmx
NextJS
React
FastAPI
Postgres