Close panel

  • Home
  • Publications
  • Who we are
  • Big Data
  • Forecasts
    Searcher

    Global | BBVA Research Geopolitical Monitor: NLP for Event Tracking and Sentiment Analysis

    Published on Thursday, November 30, 2023 | Updated on Thursday, November 30, 2023

    Big Data techniques used

    Global | BBVA Research Geopolitical Monitor: NLP for Event Tracking and Sentiment Analysis

    Summary

    The BBVA Research geopolitical monitor gathers real-time geopolitical indicators constructed from news processed with NLP, which allow for monitoring the media reaction to geopolitical and social events. This note has been published in the European Money and Finance Forum (SUERF).

    Key points

    • Key points:
    • Big Data and Geopolitics: The use of Big Data techniques, especially the processing of unstructured information and NLP models, is growing in its application to the geopolitical field due to its usefulness in modeling its complex and interconnected environment.
    • "Text to Numbers" using NLP: The GDELT database has been used, which utilizes different dictionaries to categorize information from global media news and quantify its sentiment. The result is indicators that consider both the volume and sentiment of news in specific categories, such as protest or conflict, for various countries and territories.
    • The values are smoothed and normalized in each case with the goal of providing a qualitative analysis based on the past of each indicator.
    • As an example, through the monitoring of these indicators, it has been possible to observe and monitor in real time the heterogeneous impact of the Hamas-Israel conflict depending on the theme and territory considered.

    Geographies

    Topics

    Authors

    Alvaro Ortiz BBVA Research - Head of Analysis with Big Data
    Tomasa Rodrigo BBVA Research - Lead Economist
    Jorge Sicilia BBVA Research - Chief Economist of BBVA Group

    Documents and files

    Report (PDF)

    Suerf-Note_BBVA_Research_Geopolitics_Monitor_November_2023.pdf

    English - November 30, 2023

    New comment

    Be the first to add a comment.

    Load more

    You may also be interested in