Data Science Lab

Substantive research

Applications of data science to political science and public policy

Substantive

Citizen and elite activity on Wikipedia

This project aims to study user activities on Wikipedia – by politicians on the one hand and ordinary voters on the other – by extracting insights from strategic behaviour of political elites as well as into the role of Wikipedia in digital campaigning and public attention in general
Substantive

Paying Attention to Attention

This project raises the question of whether online sources and social media platforms exacerbate polarisation and inequality in political knowledge and behaviour, and what role weak ties play in moderating citizens’ information diets.
Substantive

Lancet Countdown on Health and Climate Change

Hertie School Data Science Lab contributes to the Lancet Countdown through development and application of machine learning and natural language processing solutions to track the governance layer of health implications of climate change.

Substantive

Dynamic Election Forecasting

Combining historical information about German federal elections with current empirical data, the project created the Zweitstimme.org electoral model that accompanied the campaign for the 2017 Bundestag election with scientifically sound forecasts. 
Substantive

Territorial control in civil wars: Theory and measurement using machine learning

This article advances a new measurement strategy for territorial control in asymmetric civil wars. The strategy builds on a theoretical model of rebel tactics, by which rebels use terrorism less when they control a given area – preferring conventional tactics, which require higher levels of territorial control
Substantive

Bread Before Guns or Butter: Introducing Surplus Domestic Product (SDP)

Scholars systematically mismeasure power resources and military burdens by using gross domestic product (GDP) as a proxy for the income states can devote to arming. This paper introduces the new measure of surplus domestic product (SDP) corrects for this measurement error by decomposing subsistence income and surplus income from total GDP.