Geospatial Analysis – Sochi Olympics and Social Media

While millions of people around the world were enjoying the Olympics in Sochi on television, DigitalGlobe’s geospatial analysts were imaging the events via the company’s constellation of commercial high-resolution satellites and adding layers of social media data from web intelligence company, RecordedFuture.

[iframe src=”http://www.digitalglobe.com/interactive/sochi-tweet/” width=”100%” height=”600″]

An up-to-date imagery basemap provides important context when observing an event such as the Sochi Olympics. With such a basemap for reference, DigitalGlobe’s constellation frequently collected imagery of Sochi to monitor changes on the ground as the Olympic venues and surrounding infrastructure were constructed in preparation for the Games.

As the Olympics began, DigitalGlobe’s analysts tapped into the power of Geospatial Big Data™ – information and insight taken from imagery and derived from various sources such as social media. We particularly leveraged RecordedFuture’s open source web content to access location-tagged and sentiment-enriched social media data juxtaposed with our own high resolution imagery, in order to analyze overall activity, linguistic composition and mood for people around Sochi, including, local citizens, journalists, spectators and Olympic athletes.

GEOTAGGED EVENTS
DigitalGlobe’s analytics team used Recorded Future’s new geotagged social media data features to monitor the events in near real time. Geotagged tweets not only tell us what is happening and when it is happening, but where they are happening very precisely, often within meters of the events. The analysis of tweets from users who shared their locations provided insight about which roads visitors used for travel, popular venues that spectators visited, and the locations in which people gathered to socialize. The importance of a current imagery basemap was especially noticeable (as shown below) for mapping the Tweets against new buildings and athletic venues.

September 2005                                                                                                                    February 2014

 DIVERSITY OF LANGUAGE
Another way that social media can enhance our understanding of an area is by analyzing the languages that are used in a given location. In some places only one language is common, and any occurrence of out-of-place languages could merit further analysis.

In other places, many languages may be observed, allowing us to analyze the linguistic diversity of the area. By looking at the languages used in a particular area over time, we can identify ethnic enclaves and improve our insight into the human landscape of an area, which can further be analyzed for change over time. In the case of the Sochi Olympics, this change was dramatic.

According to Chad Shuey, a Geospatial Scientist at DigitalGlobe, “In the course of three weeks we saw the area around Sochi transform from a relatively homogeneous corner of Russia, to one of the most linguistically and culturally diverse places on the planet, and then back again.”

 

Russian vs Other Languages, before and during games (Red=Russian, Blue=Other)

  Olympic Village Area Language Diversity. orange = 1 unique language, red = many unique languages used.

SENTIMENTS AND ENRICHMENTS
Since the dawn of the Arab Spring, social media analysis has showed a focus on related sentiment. Recorded Future’s social media data is often geolocated and is enriched with additional metadata including positive and negative sentiment, and violence and protest indices, among other things. Using such sentiment information, we analyzed Sochi and the surrounding region and determined that, perhaps not surprisingly, sentiment around the Olympics was very positive but it varied in other parts of the region. This geographic analysis of sentiment revealed patterns that helped show us where to focus our attention (and satellites!).

Social Media Sentiment in the Sochi Region (Green is Positive, Purple is Negative, Gray is neutral)

One interesting location was the nearby city of Krasnodar, which is generally positive, but had negative sentiment in the run up to the games. Additional analysis of this location using Recorded Future revealed some possible explanations of this negative sentiment. These included references to the firm that was contracted to euthanize stray dogs in Sochi, which is headquartered in Krasnodar.

The analysis of information such as RecordedFuture’s geotagged social media data can provide DigitalGlobe and our customers with information about places and people that can’t be observed in imagery. This information can be used to monitor unrest, characterize the human landscape, and ultimately enhance our understanding of our changing planet.