2022
PODS-workshop: the Skagerrak sessions
Time and place: Nov. 17, 2022 9:00 AM – Nov. 18, 2022 5:00 PM, Aud 6, ES, Blindern
This workshop brings together Danish and Oslo-based political data scientists.
Each section is organized broadly by topic, each with a data-, method- and substantive-oriented presentation. Each presenter has 45 minutes to present and receive comments on the paper / presentation. Each presenter should spend no more than 20 minutes. Then the discussant will have up to 10 minutes. The remaining time, at least 15 minutes, is left for Q&A with the audience. There will be a 15 min break in between each paper.
Programme
Thursday
0900 – 1230 Governing I (data, methods, substantive)
- Solveig Bjørkholt - OSINT: Telegram
- Discussant: Philipp Lutsher
- Zoltan Fazekas – Testing campaigning/incumbency roles theory using short text classification (Twitter) & BERT
- Discussant: Philipp Broneicki
- Roman Senninger - How do people communicate their political ideas?
- Discussant: Øyvind Stiansen
1330 – 1700 Governing II (data, methods, substantive)
- Jonas Schmid - Measuring Symbolic Politics: Introducing New Data on
Buildings and Monuments from across the World- Discussant: Francesca Jensenius
- Emil Aas Stoltenberg - Regression discontinuity design with right-censored survival data
- Discussant: Jeremy Spater
- Frederik Hjorth - Losing Touch: The Rhetorical Cost of Governing
- Discussant: Jørgen Bølstad
Friday
0900 – 1230 Coordination and segregation (data, methods, substantive)
- Tore Wig - Political Science: A new dataset on scientific activity around the world at subnational scales
- Discussant: Solveig Bjørkholt
- Jeremy Spater - Measuring segregation and exposure using K-nearest-neighbors methods
- Discussant: Emil Aas Stoltenberg
- Francesca R. Jensenius - Coordination failures and the challenge of constructing a majority in Indian elections
- Discussant: Jacob Nyrup
1330 – 1700 Legislative (data, methods, substantive)
- Philipp Broniecki – European Parliament Dataset
- Discussant: Roman Senninger
- Bjørn Høyland - Strategic roll call vote requests
- Discussant: Zoltan Fazekas
- Martin Søyland - Party Conflict and Coalition Control in Parliament
- Discussant: Frederik Hjort
Political Data Science Hackathon presentations
Time and place: Aug. 4, 2022 10:00 AM – 3:30 PM, Akersveien 26 (SSB), Auditorium
We would like to invite to the final student presentations in the new summer course “Political Data Science Hackathon” (ISSSV1337)
For six weeks this summer, students have worked on various data science-related cases provided by five different work-life organizations. On Thursday the 4th of August, they will present their findings to an open audience at the auditorium in Statistics Norway.
Programme
- 10:00-10:15: Introduction by course holders
- 10:15-10:30: Introduction by ISS
- 10:30-11:00: First presentation: Statistics Norway
- 11:00-11:30: Second presentation: The Office of the Auditor General of Norway (OAG)
- 11:30-12:30: Lunch in the SSB cafeteria
- 12:30-13:00: Third presentation: OPX
- 13:00-13:30: Fourth presentation: NORAD
- 13:30-14:00: Fifth presentation: United Nations Association of Norway (UNA)
- 14:00-14:30: Deliberation in the jury
- 14:30-14:45: Announcement of 1-UP awards
- 14:45-15:00: Announcement of winners
- 15:00-15:30: Conclusion and mingling
- 15:30-17:00: PODS Celebration at Vulkan Mathallen
What is ISSSV1337?
The Political Data Science Hackathon course (ISSSV1337) was made by Alex Moltzau, Sondre Elstad and Solveig Bjørkholt. With help and inspiration from people in the PODS group, we were able to construct a course that is trying out a new way of learning. Thanks to the International Summer School at UiO, we were able to pilot our idea. What is the idea? Well...
Students work in groups with a pass/fail grade to motivate collaboration and peer learning.
The groups develop solutions for work-life relevant cases provided by different organizations, which touch upon how one could use data for social purposes.
The organizations join us in the final presentations to decide on a winner team.
During the course, we lecture a bit on data science related concepts (general programming, databases, APIs, webscraping, text handling, machine learning), but we also utilize workshop hours, which allow students to learn from each other and ask questions related to their cases.
Collecting data on legislative activities in the EP
Time and place: June 2, 2022 10:00 AM – 11:30 AM, Nydalen
Philipp Broniecki and Bjørn Høyland present their paper on Patterns of Roll-Call Requests in the European Parliament. The presentation will focus on the technical issues involved in shaping, phrasing and organizing information from a large set of semi-structured web-pages.
The presentation is based on a new paper entitled Patterns of roll-call vote requests in the European Parliament. The paper is a part of the Research Council Norway funded StREP project. The abstract reads as follow:
Studies of roll-call votes provide us with a unique insight into individual-level legislative behaviour in EU politics. However, critiques warn against generalizing from roll-calls because they are requested strategically, for example, to signal a policy position or to enforce party group cohesion. What are the patterns of roll-call vote requests and do they suggest strategic behaviour? To address this question, we leverage newly collected data on all roll-call vote requests in the European Parliament for the last two-and-a-half legislatures, from 2009 until today. We sketch the patterns of roll-call requests over time, across procedures, policy areas, and actors to evaluate the degree of (un)representativeness of roll-call votes along these dimensions. We contribute to a rich literature on legislative behaviour in the European Union that is based on roll-call voting and that assumes that roll-calls are a representative sample of all voting in the European Parliament. In addition, we add to the scholarly debate on the representativeness of roll-call votes with a more comprehensive large-scale study than has been carried out on EP legislation to date.
Elite political dynamics under competitive authoritarianism
Time: May 5, 2022 10:00 AM – 11:30 AM
Haakon Gjerløw presents his ongoing effort to collect a comprehensive dataset on the political system of Zambia since 1996.
In this seminar, Haakon Gjerløw will focus particularly on the systematization of the parliamentary minutes and attempts at measuring parliamentary unity through the analysis of these texts. Tracing organized collective action in parliamentary speeches has been successful for consolidated democracies, but it is unknown whether similar methods can be applied to competitive autocracies.
dScience
Time and place: Apr. 7, 2022 10:00 AM – 11:30 AM, Eilert Sundts hus, Auditorium 3
Ingrid Kristine Glad (dScience)
dScience – Centre for Computational and Data Science – at the University of Oslo (UiO) is an interdisciplinary centre developing and supporting research within computational science and data science across UiO and together with partners in industry and public sector.
In addition to hosting research programs and projects, dScience develops mechanisms collecting, managing and sharing high-quality data for academic and business development purposes in Norway and internationally. dScience focuses on basic, long-term research, creates collaboration across disciplines and sectors, offers community services and contributes to the education and supervision of students.
Cyber operations and the Russo-Ukrainian war
Time and place: Mar. 17, 2022 10:00 AM – 11:30 AM, Nydalen
What role are cyber operations playing in the Ukrainian war? In this presentation, Dr Philipp Lutscher will elaborate on how Ukraine became a testbed for Russian cyber actions since 2014, talk about different activities, such as Denial-of-Service attacks and their motivation, and show that such cyber operations played a rather marginal role in the conflict thus far. Finally, Philipp will talk about Russia's strategy of information control, more precisely, the use of bots, disinformation campaigns, and censorship.
The idea is to link the current developments with scientific studies on the topics out there. The talk is open to students and staff alike.
Does Politicization Make International Courts Uncomfortable? Evidence from the European Court of Justice
Time and place: Feb. 3, 2022 10:00 AM – 11:30 AM, Nydalen
Silje S. L. Hermansen presents joint work with Urska Sadl entitled Judicious judging: The effect of political debate on judicial decision making.
Authors: Silje Synnøve Lyder Hermansen and Urška Šadl
Abstract:
How do courts react to political debate? Their legitimacy hinges on their ability to uphold the law impervious to political pressure. The authority of the ruling follows from the court's reputation as apolitical. Politicization – understood as salient questions that divide stakeholders – threatens to erode this image. It imposes conflicting demands: Courts can prove their worth by solving questions for which there is a demand for a legal solution, but they risk their reputation in a politically divided environment.
Literature on judicial independence focuses on direct and explicit threats issued by the political branch. Arguments of legislative override typically predict that political division empowers the judiciary. In contrast, we argue that courts are sensitive even to the potential eruption of conflict because they value their diffuse legitimacy. Effective judicial policy making is thus a question of strategic timing.
To test our theory, we study the European Court of Justice. Known as the “dark horse of integration”, it is an independent court with political sway. Yet politicization profoundly affects its decision making. We draw on unprecedented data mapping out the legal ambit of all preliminary cases relating to the Free movement of people throughout the history. We find that the Court holds back on the legal ambit of rulings, preferring to bolster its institutional authority, in times of political division. Reversely, it broadens the ambit when salience is high and conflict low. Thus, when salience and division combine, they null each other out; leaving traces only in the decision-making process.
2021
stortingscrape: An R package for accessing data from the Norwegian parliament
Time and place: Oct. 14, 2021 10:00 AM – 11:30 AM, Seminarrom, Nydalen
Martin Søyland (Political Science, UiO) presents his R package stortingscrape. This R package aims to effectivize this process for Norwegian parliamentary data. The package makes the data easily accessible, while also being flexible enough for tailoring the different underlying data sources to ones needs. The package philosophy revolves around three core consepts: 1) simplify data formats as much as possible, 2) make interconnected sources of data easily mergable, and 3) minimize overlap in information for different retrival functions.
A wide variety of parliamentary data have been made available to the public in several countries over the last decade. Be it through front-end websites or backend APIs, researchers on parliaments have never had easier access to large amounts of data than they do now. However, both front-end and API scraped data often come in formats (.html, .xml, .json, etc) that require substantial structuring and pre-processing before they are ready for subsequent analyses.
In this paper, I present the stortingscrape package for R. Specifically, stortingscrape makes retreiving data from the Norwegian parliament (Stortinget) through their easily accessible back-end API. The data requested using the package require little to no further structuring. The scope of the package, discussed further below, ranges from information on general data on the parliament itself (rules, session info, committees, etc) to data on the parties, bibliographies of the MPs, questions, hearings, debates, votes, and more.
The main goal of stortingscrape is to allow researchers to access data from the Norwegian par- liament easily, but also still be able to structure the data according to ones need. Most importantly, the package is facilitated for weaving together different parts of the data.stortinget.no API.
High-performance computing resources at SIGMA2 and USIT – services available to UiO research and beyond
Time and place: Sep. 9, 2021 10:00 AM – 11:30 AM, Kristen Nygaards hus 2410 and Zoom
In this talk Jon Kerr Nilsen (Section Manager - Department for Research Infrastructure Services at USIT) will give an overview of which services are available for researchers at UiO when it comes to large national High Performance Computing (HPC) resources, local HPC resources as well as more lightweight and/or specialized computing services - in short, covering all computing needs requiring more than your laptop can handle.
daiR: OCR with Google Document AI in R
Time and place: June 17, 2021 10:00 AM – 11:30 AM, Zoom / Nydalen
Optical character recognition (OCR) promises to open vast bodies of historical data to scientific inquiry, but OCR can be cumbersome when documents are noisy. The past 18 months have seen the launch of new OCR processors with vastly improved accuracy. In this seminar, Thomas Hegghammer will give an overview of the latest tools and present a new R package that offers access to the most powerful of them all, Google Document AI.
The R package can be found at dair.info.
Reporting Public Opinion: How the Media Turns Boring Polls into Biased News
Time and place: May 6, 2021 10:00 AM – 11:30 AM, Zoom
Opinion polls are not reported in the media as unfiltered numbers. And some opinion polls are not reported at all. This talk by Zoltan Fazekas (Copenhagen Business School) is about how polls travel through several stages that eventually turn boring numbers into biased news.
The theoretical framework describes how and why opinion polls that are available to the public are more likely to focus on change, despite most polls showing little to no change. These dynamics are empirically demonstrated using several data sources and measurements from two different democracies (Denmark and the U.K.) covering several years of political reporting. In the end, a change narrative will be prominent in the reporting of opinion polls which contributes to what the general public sees and shares, further consolidating a picture of volatile political competition.
Violent Contention and Decolonization: Evidence from the 1919 Egyptian Revolution
Time and place: Apr. 15, 2021 10:00 AM – 11:30 AM, Zoom
Neil Ketchley presents Violent Contention and Decolonization: Evidence from the 1919 Egyptian Revolution.
Civil resistance scholars emphasize the efficacy of nonviolent protest in eliciting policy concessions. A parallel literature suggests that regimes are more likely to offer concessions following episodes of collective violence. Arbitrating this debate is difficult as we rarely have access to real-time decision making by elites. This paper takes up this question using the case of the 1919 Egyptian Revolution, when the British colonial authorities belatedly allowed Egyptian nationalists to attend the Versailles Peace Conference. To conduct its analysis, the paper draws on a corpus of newly digitized and OCR’d telegrams sent from the authorities in Cairo to the British Foreign Office. A structural topic model is used to identify telegrams calling on London to make political concessions. Analysis suggests that telegrams were more likely to contain the concessions topic following attacks on colonial communications infrastructure. Sentiment analysis and a qualitative reading of declassified colonial documents suggests a mechanism: officials feared that further attacks would irrevocably undermine the day-to-day functioning of British rule in the country.
Improved multilevel regression with post-stratification through machine learning (autoMrP)
Time and place: Feb. 18, 2021 10:00 AM – 11:30 AM, Zoom / 830
Philipp Broniecki presents joint work with Lucas Leeman and Reto Wüest on an R-package for Improved multilevel regression with post-stratification through machine learning (autoMrP).
Multilevel regression with post-stratification (MrP) has quickly become the gold standard for small area estimation. While the first MrP models did not include context-level information, current applications almost always make use of such data. When using MrP, researchers are faced with three problems: how to select features, how to specify the functional form, and how to regularize the model parameters. These problems are especially important with regard to features included at the context level. We propose a systematic approach to estimating MrP models that addresses these issues by employing a number of machine learning techniques. We illustrate our approach based on 89 items from public opinion surveys in the US and demonstrate that our approach outperforms a standard MrP model, in which the choice of context-level variables has been informed by a rich tradition of public opinion research.