Conference Schedule
March 21, 2024
16:00–16:05
Welcome to WiDS Budapest 2024!
Brigitta Németh
Junior Research Fellow ANETi Lab, Institute of Economics CERS, PhD student CUB
Andrea Sipos
Product Management Coach
Orsolya Vásárhelyi Assistant Professor, Center for Collective Learning, CIAS & Institute of Data Analytics, and Information Systems CUB
16:05–16:10
Opening Remarks
Réka Franciska Vas
Vice Rector of Education, Corvinus University of Budapest
16:10–16:30
Breaking codes, bridging gaps: Systematic errors and ethical modeling
Anna Emese Takács
Head of Data Science, BeHive Consulting
Abstract
In the sophisticated world of data science, the allure of complex statistical models often overshadows fundamental elements crucial to the field’s impact. Paramount among these are data quality, the contextual understanding of data, and the dynamics of the environments being modeled. Effective integration of data science teams with domain experts is essential in fostering efficient collaboration and developing solutions devoid of systematic errors.
In industries like finance, data science teams face significant challenges in avoiding systematic biases. Access to credit and loans, pivotal in facilitating social mobility, can have far-reaching, adverse effects if models are incorrectly developed. Traditional credit assessment models, criticized for perpetuating socio-economic inequalities, are increasingly being replaced by machine learning algorithms that consider a broader range of data, including transaction histories. This transition aims to democratise financial services and promote social mobility. Critical factors in these models include variable selection, market fit, and design. Without meticulous design and monitoring, these advanced models risk perpetuating biases, thus highlighting the need for a balance between technological innovation and ethical considerations. The keynote will delve into a case study exploring these aspects.
Data science’s responsibility extends beyond mere technical analysis; it involves a nuanced understanding of human behavior, as most data represent digital records of human actions. The keynote will emphasize the importance of inclusivity in forming data science teams. This approach is vital to facilitate discussions around design and long-term solutions, ensuring that diverse perspectives are considered in model development. By doing so, data science can transcend its technical boundaries, aligning more closely with ethical standards and societal needs.
16:35–16:50
Working in the fashion industry
Emese Pogácsás
VP of Engineering, Secret Sauce Partners
Abstract
Do you shop online? Do you shop for apparel and shoes online? Do you know your Gucci size?
Secret Sauce Partners is trying to solve this problem: whenever you’re shopping online, our tool provides suggestions (predictions) that help you find the right size for the given item you’d like to buy. Let’s take out our scuba diving equipment, put on our masks, and dive deep, deep into the world of sizing. What seems to be a relatively simple problem, but imagine it scaled up to multiple merchants, brands, categories, and countries – now you find yourself swept away by a strong current. Our approach is data-driven, and the aim is to cover catch-all marketplaces, like Walmart as well as niche high-fashion brands, like Nanushka. During the presentation, we’ll take a peek under the hood of the recommendation system, examining the difficulties of a pretty nasty category, bras. Bonus material: be prepared to see a shark casually swimming through the scene showcasing that IT and AI are still fields dominated by men.
16:55–17:10
Insightful Journeys: Transforming Medical Data into Actionable Disease Prevention Strategies for Healthy Aging
Elma Hot Debric
PostDoc researcher, Complexity Sciene Hub Vienna, Supply Chain Intelligence Institute Austria & Medical University of Vienna
Abstract
In an era where living longer is the norm, our health journey through the later years becomes a focal point of interest and concern. This project delves into the complex world of medical data, spotlighting the challenges and opportunities that longer lifespans present. We analyzed a vast dataset from Austrian hospitals covering two decades and approximately 9 million individuals; our research provides groundbreaking insights into how diseases evolve with age.
Our journey begins with multimorbidity – the coexistence of multiple health conditions in an individual. As we age, these multimorbidities become critical to our health, impacting our quality of life and the healthcare system. We have developed a novel approach to understanding these health paths, constructing a multilayered comorbidity network. This network shows how diseases are interconnected within and across various age groups.
The heart of our research lies in identifying and analyzing about 1200 disease trajectories, revealing how health conditions progress and intersect over time. For instance, our findings on sleep disorders in young males show diverging health paths in later years, some leading to more severe conditions like diabetes and movement disorders. Our result highlights the importance of early prevention and personalized health strategies.
We invite you to explore our preprint and join us in shaping a future where data-driven insights lead to better health outcomes for all.
https://arxiv.org/abs/2306.09773
Accessed for publication in NPJ Digital Medicine
17:15–17:30
What does social network analysis tell us about adolescents’ gossip?
Dorottya Kisfalusy
Senior Research Fellow, HUN-REN Centre for Social Sciences
Abstract
Does gossip bring closer individuals who have enemies in common? Or is it used to achieve consensus between friends? Do adolescents gossip about peers who have high reputations or about peers they disdain? Delve into the landscape of adolescent social dynamics as we explore the interplay between gossip, status competition, and relationships in secondary schools.
Adolescence is an important age of development when collective norms emerge, social exclusion often takes place, and competition for reputation is relatively intense. Gossip is used with increasing intentionality to interfere in these processes. Gossip is a universal social phenomenon that does not happen in a social vacuum. It is enmeshed in a fabric of positive and negative relationships that creates opportunities, constraints, and also motives to gossip. Drawing insights from empirical data in Hungarian secondary schools, this talk shows how state-of-the-art social network models help us answer theoretical questions about the motivations behind gossip.
17:35–18:05
Panel Discussion: Starting and Growing Data Science Careers
Moderator: Júlia Koltai
Senior Research Fellow at HUN-REN Centre for Social Sciences & Associate Professor at ELTE
PANELISTS
Prof. Andrea Kő
Director of the Institute of Data Analytics, and Information Systems CUB
Anna Széll
Product Growth Expert
Kriszta Kozma-Renge
Aspiring Data Scientist
Tímea Lestár
Senior Director of People Analytics, Managing Director of Arm Hungary
18:05-18:10
Closing
WiDS Budapest team
18:10-20:00
NETWORKING
Women in Data Science Budapest 