Daniela de los Santos
I'm a social data scientist.
Passionate about building data-driven tools that can be used for crafting effective public policies and tackling social and gender inequalities.
Hey there!👋 My name is
About Me
I hold an MSc in Data Science from the Barcelona School of Economics and a BSc in Sociology from the Catholic University of Uruguay. My expertise spans data-driven social research, gender and public policy analysis, machine and deep learning modeling, natural language processing, data processing and visualization, geospatial analysis, as well as scientific dissemination and teaching.
Since 2020 I have been working at the United Nations Development Programme (UNDP) Regional Hub for Latin America and the Caribbean, where I focus on leveraging data science tools for the design, implementation and evaluation of public policies aimed at reducing gender and social inequalities, with a particular emphasis on comprehensive care systems. You can check some of my publications here.
These are some of the tools and methods I work with:- Applied AI: LLM-based classification and information extraction, NLP, machine and deep learning (scikit-learn, PyTorch)
- Python: pandas, NumPy, httpx, pydantic, pytest
- R: tidyverse, Shiny, R Markdown / pagedown
- Geospatial: QGIS, ArcGIS and StoryMaps, PostGIS, OpenStreetMap, accessibility modeling. Python (GeoPandas, Shapely 2, GeoParquet, DuckDB Spatial, H3, OSMnx, r5py) and R (sf, leaflet)
- Data collection: web scraping, APIs (Google Places, Nominatim, Wikidata), ODK / KoboToolbox / QField
- Web and visualization: Leaflet, static sites and templating (Jinja2, Hugo), dashboards (Shiny, PowerBI, Tableau), accessible design (WCAG 2.1 AA)
- Automation and cloud: Git, GitHub Actions, scheduled data pipelines, AWS / Azure
- Statistics: Stata, survey design, econometrics
- AI-assisted development: agentic coding tools (e.g. Claude Code) to design, build and ship data products end to end
Experience
As a part of the Gender Team in the UNDP Regional Hub for Latin America and the Caribbean, I develop data-driven tools and methodologies in the gender, inclusive growth and social protection areas, and provide technical assistance to UNDP country offices and governments. I designed and lead the Care Georeferencing Tool, which builds territory-level diagnoses of care supply, demand and accessibility (spotting ‘care deserts’) and has been implemented in countries such as Colombia, Peru, Uruguay, Dominican Republic, El Salvador, Honduras, Guatemala and Zanzibar (Tanzania). Recently, I also led a study that used social media scraping and AI models (GPT, BERT) to analyze digital violence against women in leadership positions in eight Ibero-American countries.
As a part of the Gender and Development Department, I conducted research on gender issues, economic and social development, women’s economic autonomy, public policy, labor markets and social inequality, through quantitative and qualitative methodologies. Explore some of our publications in the link above.
During 2020-2021 I was the lecturer in charge of the Sociology undergrad course ‘Family, Gender and Generations’ in the Catholic University of Uruguay (UCU). Before that, I worked as a RA in Urban Studies and Gender and Mobility projects. Check my highlighted publication in the link.
I worked as Data and Macrodata coordinator in a Political Science research group in the University of Buenos Aires. The group studied the role of government information in guaranteeing access to women’s rights. I led a group of students that focused on improving methodologies to gather data on gender identity.
Education
2022 - 2023
MSc in Data Science for Decision Making
Barcelona School of Economics
GPA:
8.8 out of 10.0
Thesis: ‘Forecasting Global Refugee Flows: A Machine Learning Approach using Non-conventional Data’.
Grade: 10 (A+). Selected as one of the Top 3 Master’s Theses of the class of 2022-2023.
2010 - 2014
BSc in Sociology
Catholic University of Uruguay
GPA:
5.5 out of 6.0
Thesis (2020): ‘The childhood neighborhood as a center of gravity: Residential mobility of young people from Montevideo during their transition to adulthood.’
Grade: A+.
Featured Projects
Google Places
OSRM
webscraping
crowdmapping
PostGIS
QField
Care Georeferencing Tool
Big data-based tool that maps care supply, demand and accessibility to locate 'care deserts'. Implemented in Colombia, Peru, Uruguay, Dominican Republic, El Salvador, Honduras, Guatemala and Zanzibar (Tanzania).
Python
Google Places API
OpenStreetMap
r5py
GTFS
GeoCareTool (Python library)
Open-source library behind the Care Georeferencing Tool. Scrapes and cleans care facilities, models walking and public transport accessibility, and classifies care deserts by population group.
ArcGIS StoryMaps
Python
Leaflet
WCAG 2.1 AA
Care and Support Map for Persons with Disabilities (Bogotá)
Co-design, together with IDECA and Bogotá's Mayor's Office, of an accessible map of public care and support services for persons with disabilities and their caregivers. Includes an open-source demo built with Python and Leaflet.
R Shiny
Leaflet
ODK fieldwork
English / Kiswahili
Care and Support Map for Persons with Disabilities (Zanzibar)
Dashboard mapping care, health and education services for persons with disabilities in Unguja, Zanzibar, built with the Care Georeferencing Tool and hosted by the Zanzibar National Council for Persons with Disabilities.
Python
web scraping
pydantic
GitHub Actions
Jinja2
La GestorÃa Cultural
Non-profit side project that gathers Montevideo's cultural agenda (music, theater, film, dance and more) in one place. Scrapes 24 public sources three times a day, deduplicates and classifies around 700 events, and publishes a static site.
NLP
GPT models
BERT
webscraping
Digital Violence Against Women in Ibero-America
Social media scraping and AI models to identify, classify and characterize digital violence against women in leadership positions in eight countries (UNDP - SEGIB).
decision trees
gradient boosting
google trends
Forecasting Refugee Flows
A Machine Learning Approach using Non-conventional Data such as Google Trends and GDELT, in collaboration with UNHCR. Top 3 Master's Thesis (BSE, 2023).
CNN
VHR satellite imagery
pytorch
Detecting Informal Settlements using Satellite Imagery
Exercise to map informal settlements across the globe using VHR satellite imagery and Convolutional Neural Networks.
Browse publications by theme.
Here you'll find refereed papers, book chapters, reports, working papers and others.