*****Call for papers*****

* This conference is dedicated to Charles Tilly, a pioneer of the event analytic method and our great teacher.

Overview of the Conference

Social scientists have used various forms of “event data”—a catalogue of political events such as civil violence, riots, international conflicts, and popular protests—to understand chronological changes, geospatial variations, and dynamics of political phenomena of interest. With advent of new methodological innovations including natural language processing (NLP) and big data approaches, an increasing number of scholars have started to integrate these innovations into their event data analysis. One of the main objectives of this conference is to share experiences of the scholars who are specialized in one or more of the following research areas and techniques: event analysis, NLP, big data analysis, computer sciences, computational linguistics, forecasting and early warnings, data visualization, and other relevant fields. We will explore potentials and limitations of event analysis based on the innovative methods. We welcome research papers and courses addressing the following topics:


To construct a database of political, economic, and social events, scholars have used mass media such as newspapers as data source and coded manually to extract information about political events from these sources. However, a variety of techniques associated with NLP and big data approaches may revolutionize the ways of event data development.
  • Do the NLP techniques enable new ways of building event data?
  • What are the advantages and disadvantages of the event databases constructed by using some of the NLP techniques in comparison with the ones built manually?
  • What is the prospect of using the SNS such as Twitter as a source of event data? What are the potentials and problems associated with the use of the SNS in event data analysis?
  • Event data that are useful for social science research typically include the information about (1) who (2) did what (3) to whom, (4) when, (5) where, and (6) why. In other words, scholars would like to extract information about (1) Actor, (2) Action, (3) Target, (4) Time, (5) Location, and (6) Claim/Goal of event. Which of the six types of information can the new methods based on NLP and big data approaches extract reliably and which not? What are the reasons of the difficulty of extracting the information in a consistent and reliable manner?
  • If event data projects succeed in forecasting the occurrence of political violence around the world in a timely manner, the success will improve significantly the ability of policymakers and academics to prevent such disasters in the future. To make timely predictions of political disasters such as violence and civil war, it is vital to convert a massive number of news reports instantaneously on a real time basis into a data set for subsequent analyses. What kind of event data system is more desirable for such a purpose?


  • What types of analysis can be done using event data that are developed using the NLP and big data approaches?
  • What kinds of research questions can be addressed by NLP-based event data, and what kind of questions that can be addressed by manually-build event data cannot be addressed by NLP-based event data?
  • How can we explore geographical information in event data and apply GIS (geographic information systems) methods and visualize the findings?
  • How can we explore relational information in event data and apply network analytic methods to the data?
  • How can we predict civil violence using event data? The outbreak of violent conflicts often catches us by surprise. Why do people who have lived peacefully side-by-side for years suddenly start to behave violently toward each other? Why does political violence tend to concentrate in large waves, leading to tragic outcomes ranging from riots, to ethnic cleansing, genocide, and civil wars? How to develop analytic models predicting political violence theoretically and methodologically?

Our goal is to compile proceedings from the conference as a LAINAC report booklet. Each contributor is expected to develop his/her contribution for publication based on the feedback he/she will get during the sessions.


As one of the highlights of the conference, we welcome Roberto Franzosi, professor of sociology at Emory University, who will give a 5-day intensive seminar on the application of NLP to event analysis and visualization. Dr. Franzosi is one of the leading experts of the event analytic methods and has recently completed his own NLP-based event data systems. He teaches a graduate seminar on “Big/Small Data & Visualization” at Emory, and this intensive seminar is its condensed version with a special introduction of his own event analysis. The following topics will be covered (subject to change).

1. NLP (Natural Language Processing): What Is it all about?
2. Pre-processing tools (convert pdf, merge/split txt files, utf-8 compliance, document similarities)
3. Parsers, sentence splitters, lemmatizers, taggers (Postag, Deprel, NER-Named Entity Recognition)
4. The Stanford CoreNLP parser and the CoNLL table
5. Co-reference resolution (via Stanford CoreNLP and manual editing)
6. SVO (Subject-Verb-Object) extractor
7. Filtering S & O for social actors via WordNet
8. Geocoding of NER location (and date) values and GIS mapping in Google Earth Pro
9. SVOs network graphs in Gephi
10. Document annotator and extractor (via dictionary and dbpedia)

[NOTE] As for the 5-day intensive seminar by Dr. Roberto Franzosi, space is limited due to the classroom size. Priority is given to people (1) who can attend all five-day sessions, (2) who will present their work in the conference, and (3) who are undergraduate or graduate student of the University of Tokyo. If you would like to take the seminar, please submit your application through our online submission system (accessible from this website after January 21, 2020).


Conference registration: free. 
Reception party: 1,000 yen (Free for students).

  • Néstor Alvaro (Data Scientist, Deloitte Digital, Spain)
  • Doug Bond (Director, Virtual Research Associates, VRA & Harvard University, USA)
  • Jason Franz (HIIK Conflict Barometer Methodology Group, Heidelberg Institute for International Conflict Research at the Institute for Political Science, University of Heidelberg, Germany)
  • Roberto Franzosi (Emory University, USA)
  • Yoshifumi Kawasaki (The University of Tokyo, Japan)
  • Roudabeh Kishi (Research Director, Armed Conflict Location & Event Data Project, ACLED, USA)
  • Yoojin Koo (The University of Tokyo, Japan)
  • Antonio Ruiz Tinoco (Emeritus Professor, Sophia University, Japan)
  • Takeshi Wada (The University of Tokyo, Japan)
  • Seunghwan Yeo (Program Manager, Virtual Research Associates, VRA)
* Scholars and practitioners who have confirmed his/her participation as of December 23, 2019.

  • UTokyo LAINAC
  • JSPS KAKENHI, Project/Area Number 15KT0040 (Takeshi Wada, The University of Tokyo)
The conference is a part of the Global Event Data System (GEDS) Seminar Series "Big Data and Natural Language Processing in the Social Sciences & Humanities" (see the following website for more information: http://www.jp.lainac.c.u-tokyo.ac.jp/research/seminars/bdnlp)

Call for submissions

The conference will take place on March 2-6, 2020 at the University of Tokyo, Komaba Campus.


The submission deadline of abstract is February 3 (Monday), 2020 at 11:59 p.m. (Japan Standard Time).

Paper Title (15 words or less) 

Abstract (250 words or less)

How to Submit: All submissions must be made via the online submission system (accessible from this website starting from January 17, 2020). For the detail, follow the instructions on the submission system page.


The results of application will be notified by email to the applicants on February 5 (Wednesday), 2020 (Japan Standard Time).


The submission deadline of paper is February 24 (Monday), 2020 at 11:59 p.m. (Japan Standard Time).

Length and Style: Only draft/working papers that are 15-35 pages, double-spaced, (including footnotes, tables, figures, and bibliographies), that represent a well-thought-out idea or topic will be considered. Lengthier versions are more suitable for subsequent publications than for oral presentations.

How to Submit: Submission information will be sent to those who are accepted for presentation.