CONNECT@IPSDS 2017

CONNECT@IPSDS 2017

The 2nd CONNECT@IPSDS took place in Mannheim on January 27 and 28, 2017.

The Connect@IPSDS provides IPSDS students with the opportunity to meet faculty and peers in-person at the University of Mannheim.
Our students also meet survey and data science leading professionals as well as take part in some intensive workshops.

IPSDS Talks 2017

The 2nd "IPSDS Talks" took place at the University of Mannheim (B6, 26 Room A 101) on Saturday the 28th of January at 14:00-16:30.

IPSDS Talks is an event dedicated to modern data collection methods.

The event also featured talks by:

  • Piet Daas (Statistic Netherlands): "From Big Data to Official Statistics. Current Projects at Statistics Netherlands"
  • Trent Buskirk (University of Massachusetts Boston): “Does Big Data Mean Smaller Surveys? Exploring the Implications, Possibilities and Options for using Big Data in Survey Research.”
  • Stefan Bender (Bundesbank): “The Research and Data Center of the Bundesbank: Past, Present and Future - Challenges for Data Scientists”

 

Read about CONNECT@IPSDS 2016

 

Speakers


Piet Daas

From Big Data to Official Statistics. Current Projects at Statistics Netherlands

Abstract

Big data are everywhere nowadays and reflect many different aspects of our daily lives. Because of this, big data sources are very interesting from an official statistics point of view. The presentation will give an overview of the activities in the Center for Big Data Statistics at Statistics Netherlands including examples of new products and the skills and mindset needed.

Bio

Piet Daas is a senior methodologist in the Department of Corporate Services, Information Technology, and Methodology and a data scientist in the Center for Big Data Statistics of Statistics Netherlands. His work focuses on the use of secondary (nonsurvey) data for official statistical purposes, which began with the use of administrative data, and more recently has focused on studies in which Internet and other big data sources are used for official statistics. At Statistics Netherlands, he is a member of the big data core team, which oversees all big data activities of production, information technology, research, management and training. He teaches the big data component of the European Master of Official Statistics track at the University of Utrecht, is involved in the big data courses of the European Statistical Training Program, and is a member of the team organizing DataCamps (‘hackatons’) at the University of Twente. He is active in various European, United Nations and U.N. Economic Commission for Europe big data initiatives. He has a M.S. and a Ph.D. in the natural sciences with honors from the University of Nijmegen in The Netherlands.
 


Trent Buskirk

Does Big Data Mean Smaller Surveys? Exploring the Implications, Possibilities and Options for using Big Data in Survey Research

Abstract

So does Big Data mean the death of surveys as we know it?  As survey researchers we find ourselves in an era of unprecedented changes both in technology that can be used to collect data, but also in the declining rates of participation in surveys.  The abundance of “big data” may make the modern survey seem like a slower, more antiquated means of data collection that is at risk of extinction.  And while much insight can and is harnessed from big data, sometimes big data aren’t big enough to overcome systematic issues in coverage or measurement.  Certainly big data can be harvested for insights related to many topics.  However, when insights related to motivation, rationale, personal opinion or the combination thereof are desired, often big data doesn’t capture the full picture and cannot probe respondents to the extent that well designed surveys can.  The approach many have taken in this new era is to consider big data and surveys as mutually exclusive approaches to data gathering and insight generation.  In this brief talk we will explore the potential for combining big data and surveys. Specifically, we will discuss how, when and why big data may replace, reduce, restructure and refine traditional surveys.  Several examples will be provided to motivate a broader discussion aimed at asking us to rethink the purpose and potential of survey data collection in light of the enormity of big data.

Bio
Trent D. Buskirk, PhD received his Ph.D. in Statistics from Arizona State University with emphasis in Survey Sampling. Since that time Trent has developed specific expertise in Mobile and Smartphone Survey Designs and in the use of machine learning methods for developing sampling designs and adaptive survey protocols. Trent currently serves the Director of the Center for Survey Research and as a full professor in the Department of Management Science and Information Systems at the University of Massachusetts Boston. Prior to leading the Center for Survey Research, Trent was the Vice President of Statistics and Methodology at the Marketing Systems Group (MSG). Trent has also worked in research and development for the Nielsen Company and was an Associate Professor of Biostatistics at the Saint Louis University School of Public Health. His research and teaching interests include dual frame weighting for cell phone surveys, smartphone survey mode effects and applications of machine learning methods for survey design, data collection and weighting adjustment. Trent has also conducted research using both Probability-based and non-probability based panels in the context of nonresponse bias adjustments, mode effect evaluation and coverage issues. Dr. Buskirk’s research has been published in leading survey, statistics and and health related journals such as Field Methods, Journal of Royal Statistical Society, Social Science Computer Review, Journal of Official Statistics, Preventative Medicine, Cancer, Journal of Clinical Epidemiology, Survey Methods: Insights from the Field and Methods, Data and Analysis, Public Opinion Quarterly and the Journal of Survey Statistics and Methodology. Trent is currently the Past President of the Midwest Association for Public Opinion and the Associate Conference Chair of the American Association of Public Opinion Research. When Trent is not working or thinking about surveys, sampling, smartphones and research in general, you can find him playing resident prince to his two princesses or playing an action packed game of pickleball or tennis!


Stefan Bender

The Research and Data Center of the Bundesbank: Past, Present and Future - Challenges for Data Scientists

Abstract

The talk will give an overview of the data of the Bundesbank, the challenges of transforming a macro data approach to an integrated micro database, as well as the challenges of meta data documentation and record linkage. At the end, Stefan would like to discuss the G20 development of data access and data sharing.

Bio

Stefan Bender is Head of the Research Data and Service Center of the Deutsche Bundesbank. Before joining the Deutsche Bundesbank, Stefan was head of the Research Data Center of the Federal Employment Agency, Germany, at the Institute for Employment Research (IAB), where he developed one of the leading research data centers worldwide (2005-2015). From 1992 to 2005 he was senior researcher at the IAB and from 1990 to 1992 at the University of Mannheim. Among other positions, he is vice-chair of the German Data Forum (www.ratswd.de), an independent council of empirical researchers and representatives from important data producers in Germany. He was also chair of the Management Committee "European Cooperation in Science and Technology COST- Network – Comparative Analysis of Enterprise Data: Industry Dynamics, Firm Performance, and Worker Outcomes (CAED)" and Standing Committee member in the EU project "Data without Boundaries"