Models, mathematics, and data science: how to make sure we're answering the right questions
8 March 2023

Speaker

Dr Erica Thompson

Senior Policy Fellow in Ethics of Modelling and Simulation, LSE Data Science Institute.

More information.

Mathematical models underlie the interpretation of data at all levels, whether we are fitting a simple linear regression or taking advantage of the most sophisticated machine learning techniques. Ultimately the purpose of data science is not to describe data but to make use of data to inform real-world actions, so in order to understand the degree to which our models can help inform decisions, we need to understand how good they are. Dr Erica Thompson from the LSE Data Science Institute has recently published a book called Escape From Model Land, an accessible introduction to the pitfalls of constructing, calibrating and interpreting models.

This seminar will present a nuanced introduction to the use of models in data science, including how to make the most of the scenarios where we can justifiably have high confidence in model output as well as how to identify and work with those more speculative scenarios where we should expect to have lower confidence.

New Faces of Bias in Online Platforms
8 February 2023

Miss this event? A recording is available here

Speaker

The internet is fundamentally changing how we socialise, work, or gather information. The recent emergence of content serving services creates a new online ecosystem in which companies constantly compete for users' attention and use sophisticated user tracking and personalisation methods to maximise their profit.

Dr Aniko Hannak's research investigates the potential downsides of the algorithms commonly used by online platforms. Since these algorithms learn from human data, they are bound to recreate biases that are present in the real world. In this talk, Dr Hannak presents a measurement methodology developed to monitor personalisation algorithms in the context of platforms such as Google Search or online stores. Secondly, Aniko talks about recommendation and rating systems in the context of employment related platforms such as job search sites, freelancing marketplaces and online professional communities, and their danger to reinforce gender inequalities.

Aniko is an assistant professor at the University of Zurich, leading the Social Computing Group. She received her PhD from the College of Computer & Information Science at Northeastern University, where she was part of the Lazer Lab and the Algorithmic Auditing Group. Aniko’s main interest lies in computational social science, more specifically she is focusing on the co-evolution of online systems and their users. Broadly, her work investigates a variety of content serving websites such as search engines, online stores, job search sites, or freelance marketplaces and uncovers potential negative consequences of the big data algorithms that these websites deploy.

Miss this event? A recording is available here

Introducing the Online Harms Observatory
11 January 2023

Miss this event? A recording is available here

Pica Johansson

Researcher on the Online Safety Team at The Alan Turing Institute.

More information.

The Online Harms Observatory is a new analytics platform from the Alan Turing Institute’s Public Policy Programme. It combines large-scale data analysis and cutting-edge AI developed at The Turing to provide real-time insight into the scope, prevalence and dynamics of harmful content online. It aims to help policymakers, regulators, security services and civil society stakeholders better understand the landscape of online harms. Initially, it will focus on online hate, personal attacks, extremism and misinformation. The Observatory is supported by the Department for Digital, Culture, Media and Sport (DCMS).

This seminar session introduces the Observatory and will include a live demo of how it works with a case study on tracking abuse against Premier League football players on Twitter, stemming from a report commissioned in August 2022 by Ofcom.

Miss this event? A recording is available here.

Does Epistemic Vice Explain Corporate Misconduct?
19 October 2022

Miss this event? A recording is available here

Dr Marco Meyer

Principal investigator of a research group at the University of Hamburg

Organisational epistemic virtues and vices shape how concerned an organisation and its members are about the quality of the information that they circulate.

Epistemically virtuous organisations are better at generating, sharing, and storing knowledge than epistemically vicious organisations. Complementing case studies suggesting that organisational epistemic vice may drive corporate misbehavior, Dr Marco Meyer outlined how his team takes a quantitative approach to test the hypothesis whether differences in epistemic virtue and vice explain differences in corporate misconduct. This incolved analysing online employee reviews of 810 US companies between 2010 and 2020 across a broad range of sectors

Miss this event? A recording is available here.

Becoming a data scientist: what it means to put data to work
14 September 2022 

Miss this event? A recording is available here.

Professor Anne Beaulieu

Director of the Data Research Centre at the University of Groningen

The job of data scientist has been hailed as one of the most exciting emerging roles of the past decade. Who are data scientists and how do you become one? How has data-driven work led to this new role and which skills, dilemmas and achievements are associated with it? What does it mean to do that job, personally, professionally and socially?

Drawing from her recent book, Data Science and Society: A Critical Introduction with Sabina Leonelli, Professor Anne Beaulieu spoke on becoming a data scientist and on the skills that data scientists need to put data to work.

Miss this event? A recording is available here.

The Making of a French Migration Crisis
15 June 2022

Dr Michelle Reddy and Dr Hélène Thiollet

Center for International Research (CERI) at Sciences Po

Mass inflows of migrants are believed to create public anxiety and impact attitudes and votes. However, few scholars have systematically explored the construction of migration as a crisis.

In this session of the CIVICA Data Science Seminar Series, Dr Michelle Reddy and Dr Hélène Thiollet used automated text analysis on a longitudinal dataset extracted from French newspapers to locate the emergence of 'migration crisis' discourse in 2015. Absent massive inflows of migrants or refugees in France contrary to other European countries, this showcases a politicised framing of a humanitarian issue that might have created a sense of slow-burning crisis in French society.

A New Approach to Visualizing Spatial Exposure Data for Comparing Earthquakes
1 June 2022

Miss this event? A recording is available here.

Professor Stephanie Lackner

Assistant Professor in Economics at the IE Business School and the School of Global and Public Affairs at IE University

The comparison of different earthquakes is a popular tool to highlight particular aspects of one or several events. The objective is usually to demonstrate differences in the social conditions and how those affect the outcome of earthquake impacts. While earthquake magnitude is always presented in such comparisons, earthquake shaking is often not considered, even though shaking is the crucial determinant for impacts. Whenever differences in shaking are discussed, these discussions usually provide technical details that describe why the shaking was different and neglect to present how the shaking was different.

In this session of the CIVICA Data Science Seminar Series, Professor Stephanie Lackner, demonstrated an approach that utilizes two separate but complimentary steps of data visualizations which can facilitate an effective communication of earthquake shaking and population exposure data to non-experts. 

Miss this event? A recording is available here

Modeling Sustainable Development from the Bottom Up: Coupling Open Data and Agent Computing to Inform Policy Prioritisation
4 May 2022

Miss this event? A recording is available here.

Dr Omar A Guerrero

Head of Computational Social Science Research
The Alan Turing Institute

The explicit acknowledgement of the complexity of the Sustainable Development Goals (SDGs) is one of the main innovations of this international agenda. However, the formal analysis of complex systems in the SDG literature remains scant, as most of the focus is given to (top-down) aggregate models such as systems dynamics and networks of indicators.

In this talk, Omar argued that an adequate treatment of complexity needs to look at development as a bottom-up process, with macro-level outcomes emerging from micro-level interventions. From a quantitative point of view, popular methodologies such as statistical analysis and machine learning are inadequate to deal with this vertical causation since the existing data are aggregate and coarse grained (typically annual development indicators). To resolve this, models with explicit agent-level causal mechanisms are needed, and agent computing is the right tool to create them. Omar presented the research program of Policy Priority Inference, which employs agent computing to model the SDGs from the perspective of public expenditure interventions. Omar discussed several applications related to policy coherence, policy resilience, feasibility, fiscal federalism, accelerators, and bottlenecks; as well as the country-case studies in which they have been applied. This programme provides a fresh perspective to the challenges of multidimensional development, and a rigorous approach to exploit not only indicators, but also new sources such as open spending data.

Miss this event? A recording is available here.

Internet Communities and the French Presidential Election
20 April 2022

Miss this event? Watch a recording here.

Professor David Chavalarias

Director of the Complex Systems Institute of Paris Ile-de-France and Vice-President of the Complex Systems Society.

From election campaigns to the war in Ukraine, social networks are now used on a massive scale to influence public opinion. In this new session of the CIVICA Data Science Seminar Series, Professor David Chavalarias, Director of the Politoscope project and author of a newly released French-language book entitled Toxic Data, explains how online advocacy groups can shape the course of events in a country, even from outside its borders.

Miss this event? Watch a recording here.

The Science of Success: Quantifying Outcomes in Social Systems
9 March 2022

Miss this event? Watch a recording here.

Professor Albert-László Barabási

Professor of Network Science at Northeastern University and Director of the Center for Complex Network Research

Our current approach to success is driven by the belief that predicting exceptional impact requires us to detect extraordinary ability.

Despite the long-standing interest in the problem, even experts remain notoriously bad at predicting long-term impact. Success becomes suddenly predictable, however, if we see it not as an individual but a collective phenomenon: for something to be successful, it is not enough to be novel or appealing, but we all must agree that it is worthy of praise.

If we accept the collective nature of success, its signatures can be uncovered from the many pieces of data around us using the tools of network and data science.

In this talk, Professor Albert-László Barabási focused on a general, quantitative theory of success and signal the emergence of a new research field - The Science of Success.

Miss this event? Watch a recording here.

Embedding Regression: Models for Context-Specific Description and Inference
23 February 2022

Miss this event? Watch a recording here.

Professor Arthur Spirling

Professor of Politics and Data Science at New York University

"Political scientists commonly seek to make statements about how a word’s use and meaning varies over circumstances—whether that be time, partisan identity, or some other document-level covariate. A promising avenue is the use of domain-specific word embeddings, that simultaneously allow for statements of uncertainty and statistical inference."

In this session, Professor Arthur Spirling, introduced the a la Carte on Text (ConText) embedding regression model for this purpose and evaluate how this method outperforms well-known competitors for studying changes in meaning of words across groups and time.

Miss this event? Watch a recording here.

Information and Irregular Migration: Evidence from a Field Experiment in Nigeria
9 February 2022

Dr Alexandra Scacco

Senior Research Fellow in the Institutions and Political Inequality unit at WZB

Policy projections and recent research suggest that large numbers of migrants from sub-Saharan Africa will continue to attempt to make their way to Europe over the next few decades. In response, European countries have made and continue to make significant investments in information campaigns designed to discourage irregular African migration. 

Dr Alexandra Scacco from WZB Berlin led us through her team's research and field experiment on the effects of having detailed, accurate information about features of the journey along the “Mediterranean route”. Alex explored migration-related risks, interest in attempting irregular migration, and actual migration from households in Nigeria, helping to shed some light on why and how certain groups of people choose to embark on this migration journey.

Understanding Social Media censorship with Text Matching
26 January 2022

Miss this event? Watch a recording here.

Professor Margaret Roberts

Associate Professor in the Department of Political Science at the University of California, San Diego and co-director of the China Data Lab at the 21st Century China Center.

 
"Social media users in China are censored every day, but it is largely unknown how the experience of being censored affects their future online experience. Are social media users who are censored for the first time flagged by censors for increased scrutiny in the future? Is censorship 'targeted' and 'customised' toward specific users? Do social media users avoid writing after being censored? Do they continue to write on sensitive topics or do they avoid them?"

This is the opening to the latest research on Adjusting for Confounding with Text Matching by Professor Margaret Roberts and her research team, Brandon Stewart and Rich Nielsen.

In this session, Margaret shared how the method of text matching can help to address the problem of confounding in observational studies which ultimately supports credible causal inferences about the effects of experiencing censorship. She also validated her research approach and illustrated the importance of conditioning on text to address confounding with two applications: the effect of perceptions of author gender on citation counts in the international relations literature and the effects of censorship on Chinese social media users.

Miss this event? Watch a recording here.

What is Feminist Data Science?
12 January 2022

Miss this event? Watch a recording here.

Speaker

Professor Lauren Klein

Associate Professor in the Departments of English and Quantitative Theory and Methods at Emory University

What is feminist data science? How can feminist thinking be incorporated into data-driven work? How are social scientists bringing together data science and feminist theory in their research?

Drawing from her recent book, Data Feminism (MIT Press, 2020), Professor Lauren Klein presented a set of principles for doing data science that are informed by the past several decades of intersectional feminist activism and critical thought.

Miss this event? Watch a recording here.

More than Words: How Political Rhetoric Shapes Voters’ Affect and Evaluation
3 November 2021

Miss this event? Watch a recording here.

Speaker

The study of political campaign in academia so far has mainly focused on the text of speech and the way that candidates communicate to their audience.

In this session, Professor Christopher Lucas led us through his team's research on how we are ignoring another tremendously important component of how voters assess a potential political leader - the non-textual elements and political rhetoric.

Miss this event? Watch a recording here.

Serendipity or confinement? Deconstructing the principles, effects, and uses of algorithmic recommendation
20 October 2021

Miss this event? Watch a recording here.

Speaker

Professor Camille Roth

Associate Research Professor at French National Centre for Scientific Research

 

The effects of algorithmic recommendation in online systems is a subject of growing interest and, also, of sometimes conflicting results, depending on whether it is said to contribute to expand or to restrain the horizon and serendipity of users.

In this session Professor Camille Roth explored his team's research on the interplay of influences between users' preferences and content suggestion from recommendation systems. Camille discussed online platforms such as Youtube and Deezer to highlight that sometimes, the story of the internet is not as simple as it seems.

Miss this event? Watch a recording here.

Framing A Protest: Determinants and Effects of Visual Frames
6 October 2021

Miss this event? Watch a recording here.

Speaker

Professor Michelle Torres

Assistant Professor at Rice University

The information that media provides to citizens fuels their attitudes and opinions towards social movements but existing analyses of the ways in which media portrays protests have overlooked the impact of visual images.

Professor Michelle Torres guided us through her research on how visual framing of protests in media outlets can exert influence on citizens' political attitudes and participation, proving once again, that a picture is indeed worth a thousand words.

Miss this event? Watch a recording here.

Understanding Beautiful Places and Wellbeing with AI
22 September 2021

Miss this event? Watch a recording here.

Speaker

Professor Suzy Moat described how millions of ratings from an online game called ‘Scenic-or-Not’ and a mobile app called ‘Mappiness’ has begun to offer new answers to the age-old question of whether beautiful environments are good for our health and happiness.

This event explored how deep learning can help us understand whether beautiful places are simply natural places - or whether humans might be able to build beautiful places too.

Miss this event? Watch a recording here.

Incentives and COVID-19 vaccination uptake
8 September 2021

Miss this event? Watch a recording here.

Speaker

Professor Macartan Humphreys

Professor of Political Science at Columbia University and director of the Institutions and Political Inequality group at WZB Berlin

Recent evidence suggests that vaccination hesitancy is too high in many countries to sustainably contain COVID-19.

In this event, Professor Macartan Humphreys outlined the findings of a factorial survey experiment administered to 20,500 online respondents in Germany. With this he was able to assess the effectiveness of three strategies to increase vaccine uptake: providing freedoms, financial remuneration, and vaccination at local doctors.

Miss this event? Watch a recording here.

The Art of Quantitative Editing
2 June 2021

Miss this event? Watch a recording here.

Speaker

Dr Laura Bronner

Quantitative Editor at FiveThirtyEight

As data science becomes an ever more popular tool to support decision-making, it is important to create and implement a robust review process for the latest findings. This is the job of the quantitative editor - a new and exciting career path at the intersection of data and journalism.

In this seminar Dr Laura Bronner guides us through her experience of working as a quantitative editor at FiveThirtyEight, and how she uses her expertise to evaluate the data and code from research and articles that will go on to inform the large audience of FiveThirtyEight.

Miss this event? Watch a recording here.

Breaking the Social Media Prism
19 May 2021

Miss this event? Watch a recording here.

Speaker

Professor Chris Bail

Professor of Sociology and Public Policy at Duke University

In an era of increasing social isolation, platforms like Facebook and Twitter are among the most important tools we have to understand each other. We use social media as a mirror to decipher our place in society but, it functions more like a prism that distorts our identities, empowers status-seeking extremists, and renders moderates all but invisible.

In this seminar Professor Chris Bail explores the minds of these online extremists and offers innovative solutions to counter this political tribalism.

Miss this event? Watch a recording here.

Using Public Video Cameras to Detect Racial Distancing on City Streets
5 May 2021

Miss this event? Watch a recording here.

Speaker

Dr Melissa Sands

Assistant Professor of Political Science at the University of California, Merced

Do people consciously or unconsciously avoid others of certain races on city streets? Is casual "everyday racism" present just on a simple walk in the neighborhood?

In this seminar Dr Melissa Sands explored how, using publicly-available traffic camera feeds in combination with a real-world field experiment, pedestrians behave in the presence of other racial groups and discovered an implicit bias against certain people of colour. 

15:00 BST - Welcome Introduction
Dr. Erica Thompson, Senior Policy Fellow, LSE Data Science Institute 
Setting the scene: Brief intro to the speaker and her research

15:05 BST - Seminar Session
Dr. Melissa Sands, University of California, Merced
"Seeing Racial Avoidance on City Streets"

15:45 BST - Research Discussion
Q&A / Discussion of the research

16:00 BST - Announcement
Upcoming seminar in the series and other announcements

Miss this event? Watch a recording here.

How to detect fake news before it is written
21 April 2021

Miss this event? Watch a recording here.

Speaker

Dr Preslav Nakov

Principal Scientist at the Qatar Computing Research Institute, Hamad Bin Khalifa University

Given the recent proliferation of disinformation online, there has been growing research interest in manual and automatic fact-checking initiatives. A different direction is to analyse entire news outlets in advance.

In this seminar Dr Preslav Nakov explained the Tanbih news aggregator and COVID-19 disinformation detector platforms, which aimimg to limit the impact of "fake news", propaganda and media bias by making users aware of what they are reading.

Miss this event? Watch a recording here.

Negotiating with AI: Fairness in the Labor Market
7 April 2021

Miss this event? Watch a recording here.

Speakers

Artificial Intelligence and Machine Learning systems are increasingly being deployed to automate the process of hiring. Is it really up to the task?

In this event Professor Christo Wilson explored audits of these algorithms and services, digging into their source code to determine whether they truly contain safeguards to prevent human errors and maintaining fairness. 

Miss this event? Watch a recording here.

Tracking COVID-19 with the Financial Times
24 March 2021

Miss this event? Watch a recording here.

Speaker

One of the most impactful ways that data science can change public policy and community is the building of tools that help people discover for themselves how the world works. 

In this session John Burn-Murdoch, the key architect behind the Financial Times COVID-19 data visualisation tool shared his experience in data collection and presentation to the general public.

Miss this event? Watch a recording here.

Can Police Diversity Improve Police-Civilian Interactions?
17 March 2021

Miss this event? Watch a recording here.

Speaker

AGENDA

15:00 GMT - Welcome Introduction
Prof. Slava Jankin, Hertie School Data Science Lab
Setting the scene: brief intro to the speaker and his research

15:05 GMT - Seminar Session
Roman Rivera, Columbia University
'The Role of Officer Race and Gender in Police-Civilian Interactions in Chicago' 

15:35 GMT - Research Discussion
Q&A / Discussion on the research

16:00 GMT - Announcement
Upcoming seminars in the series and other announcements

Miss this event? Watch a recording here.

Data Science in the Time of Coronavirus and What Happens After
24 February 2021

Watch a recording of this event here

The launch event featured a round table discussion on 'Data Science and Digital Transformation in the Time of Coronavirus, and What Happens After'. Four notable experts from academia, government and business gave their perspectives on how social data science is being used to tackly the problems of the twenty-first century, and their thoughts on how the rapid transformations we have seen over the last year will result in lasting changes. 

The session will be chaired by Professor Slava Jankin, Director of the Hertie School Data Science Lab, and Professor Ken Benoit, Director of the LSE Data Science Institute. 

Hertie School and LSE are among Europe's leading social science research institutions, brought together by the CIVICA research network which will be funding ongoing programmes in data science for social applications. 

CHAIRS

SPEAKERS

Watch a recording of this event here