SICSS-ODISSEI Summer School on Computational Social Science

Introduction

Key terms: quantitative research, computational science, machine learning, network analysis, analysis of large-scale and complex linked data, Python software, advanced course, relevant for students in any PhD phase.

ECTS: 5
Number of sessions: 10
Hours per sessions: 8

This summer school aims to bring together PhD students, postdoctoral researchers, and early career faculty who are interested in computational social science. Computational social science forms an interdisciplinary field that uses algorithms and large datasets to study social phenomena and human behaviour.

Participants will be working with highly enriched commercial data on firms and economic networks. During the summer school, we will cover advanced quantitative research methods, including network analysis and machine learning, and also discuss ethical aspects of computational research.

There will be ample opportunities for students to discuss their ideas and research with the organisers, other participants, and visiting speakers. Because we are committed to open and reproducible research, all materials created by faculty and students for the Summer Institute will be released open source.

The curriculum of the summer school will be supported by ODISSEI partner organisations such as SURF (the Netherlands High Performance Computing Center), the Netherlands eScience Center, and Firmbackbone.

 

Entry level and relevance

Good knowledge of object-oriented programming such as R or Python is required before the course, but specific language or software knowledge is not required. Please contact the lecturer in case you doubt whether you have the required entry level.

The course is useful for students in any PhD phase as well as for post-docs and early career researchers generally.

The targeted fields of research are principally sociology, psychology, economics and political science, but researchers actively involved or interested in related fields are also encouraged to apply for the programme.

Relations with other courses

There is no distinct overlap between this course and other courses offered by the EGSH. This course teaches other or more advanced quantitative methods than the ones that are offered in other EGSH courses.

To prepare for the use of software in this course you may follow the EGSH course Data literacy through R, but this is not mandatory or necessary.

Sessions and preparations

No preparation is needed for any of the sessions.

Day 1: Introduction to ODISSEI, open science and the theme of the summer school.
During the first day you are extensively onboarded into the ODISSEI infrastructure and we set you up with a secure environment, test your python and R installations and arrange any other elements of your workflow that are needed.

Day 2: Causal inference (by ODISSEI SoDa team)
During this workshop you will explore the use of various causal inference methods in the context of large scale administrative data such as CBS and Firmbackbone.

Day 3 and 4:  Machine Learning in Python with scikit-learn (by the Netherlands eScience Center) 
During this course we take you through the basics of machine learning including main concepts, standard workflows and how to use the scikit learn package in Python. We do this using worked examples from the FirmBackBone data.

Day 5: Annotation workshop (by ODISSEI SoDa team)
In this workshop you will be shown how to set-up and evaluate a prompt workflow for an LLM and effective prompt engineering. In this workshop you will use web scraped data that can be linked to individual firms and can be annotated to understand companies’ public pronouncements.

Day 6 to 10: Research project
During the second week you will work consistently in a group on a structured research project for which you will be provided with the required infrastructure. Attendance is compulsory for the entire second week.