Global Database Of Events, Language, And Tone Research Articles

Spatial Big Data Analysis of Political Risks along the Belt and Road

As many regions along the Belt and Road have long been struggling with terrorist attacks, crimes, wars, and corruption, political risks pose important challenges for infrastructure projects and transnational investment. The objective of the article is to contribute to the identification of different types of political risks along the Silk Road Economic Belt and 21st Century Maritime Silk Road, and the visualization of their micro-level spatial distribution based on the Global Database of Events, Language, and Tone (GDELT) datasets from October 2013 to May 2018. By adopting the bivariate Moran’s I model to compare the distribution of political risks along the Belt and Road and that of the Chinese Belt and Road Initiative (BRI) investment and construction projects based on data from the China Global Investment Tracker (CGIT), the article also generates an overall political risk profile for Chinese BRI projects. Our findings show that a particularly high percentage of Chinese BRI projects are distributed in regions with high political risks. This research has important implications for the discussion and study of the BRI. First, by combining geographic spatial statistical analysis and political science conceptual frameworks, we point out the necessity to query the BRI from interdisciplinary perspectives grounded in empirical research. Second, the research delivers to researchers, academics, practitioners, consultants and policy makers interested in the BRI the latest insights into the risks and challenges along the Belt and Road. Third, it advocates policies and strategies conducive to identifying, assessing and mitigating political risks in investment along the Belt and Road and to strengthening the sustainable development of the BRI.

Embedding Models for Episodic Knowledge Graphs

In recent years a number of large-scale triple-oriented knowledge graphs have been generated and various models have been proposed to perform learning in those graphs. Most knowledge graphs are static and reflect the world in its current state. In reality, of course, the state of the world is changing: a healthy person becomes diagnosed with a disease and a new president is inaugurated. In this paper, we extend models for static knowledge graphs to temporal knowledge graphs. This enables us to store episodic data and to generalize to new facts (inductive learning). We generalize leading learning models for static knowledge graphs (i.e., Tucker, RESCAL, HolE, ComplEx, DistMult) to temporal knowledge graphs. In particular, we introduce a new tensor model, ConT, with superior generalization performance. The performances of all proposed models are analyzed on two different datasets: the Global Database of Events, Language, and Tone (GDELT) and the database for Integrated Conflict Early Warning System (ICEWS). We argue that temporal knowledge graph embeddings might be models also for cognitive episodic memory (facts we remember and can recollect) and that a semantic memory (current facts we know) can be generated from episodic memory by a marginalization operation. We validate this episodic-to-semantic projection hypothesis with the ICEWS dataset.

Embedding models for episodic knowledge graphs

In recent years a number of large-scale triple-oriented knowledge graphs have been generated and various models have been proposed to perform learning in those graphs. Most knowledge graphs are static and reflect the world in its current state. In reality, of course, the state of the world is changing: a healthy person becomes diagnosed with a disease and a new president is inaugurated. In this paper, we extend models for static knowledge graphs to temporal knowledge graphs. This enables us to store episodic data and to generalize to new facts (inductive learning). We generalize leading learning models for static knowledge graphs (i.e., Tucker, RESCAL, HolE, ComplEx, DistMult) to temporal knowledge graphs. In particular, we introduce a new tensor model, ConT, with superior generalization performance. The performances of all proposed models are analyzed on two different datasets: the Global Database of Events, Language, and Tone (GDELT) and the database for Integrated Conflict Early Warning System (ICEWS). We argue that temporal knowledge graph embeddings might be models also for cognitive episodic memory (facts we remember and can recollect) and that a semantic memory (current facts we know) can be generated from episodic memory by a marginalization operation. We validate this episodic-to-semantic projection hypothesis with the ICEWS dataset.

News Media Coverage of Refugees in 2016: A GDELT Case Study

For many citizens, the main source of information about world events is the news media. Undeniably, media coverage of topics influences the perceived importance of these topics. Recent media discourse has given extensive attention to the topic of refugees and migration. In this context, we conduct a study on news media coverage of refugees for the time period of January 1st 2016 to December 31st 2016. We rely on the Global Database of Events, Language, and Tone (GDELT) project that monitors and analyses news articles around the world. In a first phase, we explore the global media conversation on refugees along two important dimensions: article quantity and sentiment. We identify and characterize events that generated an extensive media coverage and also extreme sentiment in the news reports. In a second phase, we refine the analysis by focusing on the news media coverage related to refugees in Europe. GDELT features allow us to identify the key countries and time evolution of the media coverage. Lastly, we determine the main actors linked to refugees in the news and study their interaction through a network analysis.

Using GDELT Data to Evaluate the Con dence on the Spanish Government Energy Policy

The growing demand for affordable, reliable, domestically sourced, and low-carbon electricity is a matter of concern and it is driven by several causes including public policy priorities. Policy objectives and new technologies are changing wholesale market design. The analysis of different aspects of energy markets is increasingly on the agendas of academics, firms� managers or policy makers. Some concerns are global and are related to the evolution of climate change phenomena. Others are regional or national and they strongly appear in countries like Spain with a high dependence on foreign energy sources and high potential of domestic renewable energy sources. We can find a relevant case in Spanish solar energy policy. A series of regulatory reforms since 2010 reduce revenues to existing renewable power generators and they end up the previous system of support to new renewable generation. This policy change has altered the composition of the energy market affecting investment decisions. In this paper, we analyze the public opinion about energy policy of the Spanish Government using the Global Database of Events, Language, and Tone (GDELT). The GDELT Project consists of over a quarter-billion event records in over 300 categories covering the entire world from 1979 to present, along with a massive network diagram connecting every person, organization, location, and theme to this event database. Our aim is to build sentiment indicators arising from this source of information and, in a final step, evaluate if positive and negative indexes have any effect on the evolution of key market variables as prices and demand.

Using machine-coded event data for the micro-level study of political violence

Machine-coded datasets likely represent the future of event data analysis. We assess the use of one of these datasets—Global Database of Events, Language and Tone (GDELT)—for the micro-level study of political violence by comparing it to two hand-coded conflict event datasets. Our findings indicate that GDELT should be used with caution for geo-spatial analyses at the subnational level: its overall correlation with hand-coded data is mediocre, and at the local level major issues of geographic bias exist in how events are reported. Overall, our findings suggest that due to these issues, researchers studying local conflict processes may want to wait for a more reliable geocoding method before relying too heavily on this set of machine-coded data.