Introduction
In a two week programme, ODISSEI hosts a summer school in computational social science in the Netherlands. The summer school covers various techniques and topics that are relevant for innovative statistical research, including network analysis, how to analyse large scale administrative data, machine learning, and ethics.
The methods covered within the course will enable participants to answer questions about complex social systems and processes using rich administrative data. These methods are particularly useful in answering questions about contextual effects, group dynamics, and long-term processes. The course will also provide an introduction to using linked administrative data and open science workflow which will enable participants to use rich data for a wide range of social research topics.
ODISSEI is the Dutch national infrastructure for social science. Providing analytical tools and software, secure supercomputing power and expertise, ODISSEI is building an overarching and interlinked infrastructure to guide researchers through the rich datasets that are available in the Netherlands. SICSS-ODISSEI participants will be able to share their research, discuss their ideas and connect with other researchers in related fields, across the world.
Course information
ECTS: 5
Number of sessions: 10
Hours per sessions: 8
Key Facts & Figures
- Type
- Course
- Instruction language
- English
What will you achieve?
- After the course you will have knowledge of computational social science techniques such as Network Analysis, Machine Learning and Benchmarking.
- After the course you will have knowledge of CBS (Statistics Netherlands) administrative data, the full potential of these data, high performance computing and intensively linked data.
- After the course you will have developed your research ideas into a larger coherent project and learned how the ODISSEI infrastructure can be adapted to your own needs.
Start dates
Week 1
17-21 June 2024
Week 2
24-28 June 2024
Aims and working method
The program will involve lectures, group problem sets, and interactive work sessions. There will also be outside speakers who conduct computational social science research in a variety of settings, such as academia, industry, and government.
The summer school will offer unique access to the ODISSEI infrastructure, including administrative data at CBS, the support of Research Software Engineers (RSE) at eScience and access to the nationally representative LISS Panel. Because we are committed to open and reproducible research, all materials created by faculty and students for the summer school will be released open source.
Entry level
- The summer school is relevant for both social scientists (broadly conceived) and data scientists (broadly conceived).
- The primary target audience is postdoctoral researchers, PhD students and advanced Masters’ students.
- The targeted fields of research are principally sociology, psychology, economics and political science. However, individuals or junior researchers actively involved or interested in related fields may apply for the programme.
- No specific technical knowledge is required in advance, but participants not familiar with Python and R in advance will be provided with introductory material to acquire sufficient knowledge to participate in the course.
How to prepare
Before the summer school starts, participants should complete some readings and (if needed) a coding bootcamp, and prepare a computing environment. More information about these preparations will be sent in due time.
Session overview
On the first day there will be an introduction to ODISSEI and its various elements including Microdata Services at Statistics Netherlands, the ODISSEI Secure Supercomputer, the LISS Panel, the ODISSEI portal and the various data collections that ODISSEI supports. The day will include a range of practical, hands-on exercises using the various services, getting participants acquainted with the Dutch data landscape, and how ODISSEI supports computational social science.
Speaker: Dr. Tom Emery
ODISSEI brings together multiple data sources (administrative data, survey data, web data, etc) which allows researchers to construct multi-dimensional data which is often best analysed using Machine Learning techniques rather than traditional social science research methods. The data at CBS for example contains more than 10,000 variables so specifying the right model can be laborious. On this first day, participants will be introduced to Machine Learning and how it relates to traditional social research methods by experts from the eScience Center. The day will be interspersed with presentations by social scientists who have utilised Machine Learning in social research and hands-on experiments using data from ODISSEI.
Speaker: Expert from eScience Center
On the second day of machine learning participants will be introduced to more advanced concepts in machine learning including deep learning and neural networks with an increased focus on worked examples that participants will be able to utilise in the second week of the program. Worked examples will show how to deploy machine learning approaches in both Python and R, how to troubleshoot problems, and evaluate models.
Speaker: Expert from eScience center
Because ODISSEI enables intensive data linkage, network approaches are particularly useful when using ODISSEI infrastructure. This day introduces concepts and tools in network science. The objective of the day is that participants acquire hands-on knowledge on how to analyse social networks. Participants will be able to understand when a network approach is useful, understand the differences and similarities between a Complex Networks and a Social Network Analysis approach, describe network characteristics, and infer edges or node attributes.
Speaker: Dr. Javier Garcia-Bernardo
On the last day of the first week, a benchmarking workshop will be held, during which the students will be introduced to the concept of benchmarking and how this could be used in Social Sciences. The second part of the workshop will consist of a hands-on exercise, in which students will be given a simple prediction task (i.e., predicting a life outcome based on publicly available survey data) and will see how their proposed methods can be compared and assessed given a set of pre-defined matrices and criteria.
Speaker: dr. Paulina Pankowska
In the second week of the summer school, participants will be able to make full use of the ODISSEI infrastructure through a full scale benchmarking challenge, during which they will be put into teams and challenged to work on and submit solutions to a real-life prediction problem that social scientists and policy makers are grappling with. An example topic would be to predict educational or labour market outcomes. For an example of such a social science benchmark that was organised a few years ago see the Fragile Families Challenge in the onboarding reading.
Course instructors
Dr. Tom Emery is the Deputy Director of ODISSEI, where he is responsible for the strategic development of the infrastructure and international collaborations. Emery is an Associate Professor in the Department of Public Administration and Sociology of Erasmus University Rotterdam. Before that, he was the Deputy Director of the Generations and Gender Programme (GGP) at the Netherlands Interdisciplinary Demographic Institute in The Hague. Emery gained a PhD in Social Policy from the University of Edinburgh in 2014 and his thesis examined the interaction between financial support between elderly parents and their adult children in a number of European countries. His research also covers questions of comparative survey methodology and policy measurements in multilevel contextsEmail address
dr. Paulina PankowskaPaulina Pankowska is a postdoctoral researcher at the Sociology and Communication Science Departments of the Vrije Universiteit Amsterdam. Her research focuses on the topics of data and methods quality. She is currently involved in an ERC project investigating the problem of measurement error in the context of career and employment trajectories. She is also the task leader of the ODISSEI benchmarking task, which aims to organize an algorithm benchmark for the social sciences. The overarching goal of this project is to guide social science research towards a culture wherein different methods and techniques that are used to solve a specific problem are compared and evaluated objectively.Email address
Prof. dr. Daniel OberskiProf. dr. Daniel Oberski is a full professor of health and social data science, with a joint appointment at Utrecht University’s Department of Methodology & Statistics, and the department of Biostatistics and Data Science at the Julius Center, University Medical Center Utrecht (UMCU). Oberski was awarded with a Veni grant from NWO in 2019 . Oberski is a member of the Young Academy of the Dutch Royal Academy of Arts and Science and was also a visiting professor at the Joint Program for Survey Methodology (JPSM) at the University of Maryland, teaching experimental design.Email address
Dr. Javier Garcia-BernardoDr. Javier Garcia-Bernardo is an assistant professor at Utrecht University in the Social Data Science (SoDa) team. Before that, he was a postdoc at the University of Amsterdam and at Charles University (CORPTAX), and a data scientist at the Tax Justice Network. In his research he applies computational models to understand social and economical systems. He completed his PhD in Political Economy at the CORPNET group (University of Amsterdam), and his MSc in Computer Science at the University of Vermont.Email address
Contact
Contact and information: sicss@odissei-data.nl
Facts & Figures
- Fee
free
- Tax
- Not applicable
- Offered by
- Erasmus Graduate School of Social Sciences and the Humanities
- Course type
- Course
- Instruction language
- English