Observing Global Health and Well-being

Guest post by Marko Grobelnik, Quintelligence

Observing anything on the global scale is generally not easy, mainly due the scale of the possible data, speed of changing the situation, and resolution ranging from macro to micro. With today’s big-data and networked internet technology many things became possible, but still are far from easy. Although the data might be available and accessible, most of the areas lack standards which is causing additional problems for collecting and aggregating data into meaningful messages.

In this blog we will touch the area of real-time global media monitoring and how media can contribute to the better picture and overview to understand health and more general well-being of any part of the connected world (nowadays, most of the world).

Dealing with media (written online media in particular) has several issues. One is a lack of common standards how the content gets available, as soon as it appears. Next, being global, means dealing with many, possibly hundreds of languages. Since today’s language technologies can deal only with words and sentences, there is a big gap how to move from simple textual representation towards semantic representation, where we would want to observe the semantic and conceptual aspect of the textual content and not just lexical (words and phrases) and syntactical (sentences).

An example of such global media monitor is the system “Event Registry”, available from <http://eventregistry.org/> which track over 100,000 global media sources in near-real-time (with few minutes delay), operates across 100 languages, and aggregates global media content in a semantically meaningful way. It collects over 300,000 news articles on average per day and arranges them into events (i.e., news stories in different languages, talking about the same topic, are clustered into event clusters), events are further connected into story-lines which enable tracking of evolving topics.

Since “Event Registry” covers most of the global media reporting on any topic, it can be used also to track topics like health and well-being on different levels of resolution – from small local issues, up-to the higher level country issues and global trends.

In this text we want to show three aspects of how “Event Registry” could be used for health monitoring.

(1) Real-time monitoring of the global health, medical and well-being issues can be observed through a graphical dashboard showing just issues related to health. The health dashboard is accessible through http://eventregistry.org/currentActivity?conceptUri=http:%2F%2Fen.wikipedia.org%2Fwiki%2FHealth. It shows the incoming health related media content published somewhere in the world allowing to have an overview of what is happening in the world right now. The image below shows a snapshot of the dashboard with health related events on July 20th 2017.

(2) Another view to the health related issues is historical aggregation of a particular topic. In the example below we show the evolution of Ebola virus and reporting after 2014 till 2017. With a simple query “Ebola virus disease” the system shows over 20,000 events related to Ebola appearing after 2014. The content (over 200,000 news articles) could be viewed through a temporal intensity, geographical spread, topical spread and others. The following pictures show temporal (with the peak in October 2014) and geographical spread (West Africa and US) of Ebola related events.

(3) The third view is the micro view of a particular health related event happening somewhere in the world. As the example we selected “lead poisoning” in Nigeria on May 12th 2015 which was mostly unnoticed on the global level. With a simple query to the system, we can extract all the reports related to that event, and check the event in the context of other related events. The following two screenshots show the event itself and the storyline developed after the event.

To conclude. We presented a convenient way how to observe global health monitoring by using “Event Registry” system. The key feature is to be able to observe health issues across many languages and in the long tail, where most of the other systems have difficulties.