Citizen perception of nature on social media

Organisator Laboratory of Geo-information Science and Remote Sensing

do 14 december 2017 11:40 tot 12:10

Locatie Lumen, building number 100
Droevendaalsesteeg 3a
6708 PB Wageningen
+31 317 481 700
Zaal/kamer 1

By Maartje Holtslag (the Netherlands)

With the internet becoming available to the wide public, social media platforms have become an important part of people’s everyday life worldwide. This has drawn interest in the academic world, using the possibility to extract and analyse people’s opinions from social media. This research investigates whether data from social media can be used to determine citizen perception. A case study is carried out to find the citizen perception of nature, to assess cultural ecosystem services in the Pike-San Isabel National Forest, Colorado USA. The results are compared with the results from a study in the same area, using survey data to determine the social value of the park, in order to find whether the proposed method can replace the use of surveys. A large dataset has been created by harvesting data from different social media platforms. Relevant data from this dataset was selected using a Naive Bayes classification combined with Expectation Maximization. A sentiment analysis has been applied to extract the citizen perception from the messages. This data has been extrapolated towards social value maps using the Social Values for Ecosystem Services (SolVES) tool. After which a hotspot analysis was performed to identify hotspots with a high social value in the area. When these results were compared to the results from the previous study using surveys, it could be concluded that social media data can be used as a complement to this data to increase the value of knowledge about cultural ecosystem services.

Keywords: Cultural ecosystem services; Expectation Maximization; Hotspot analysis; Naive Bayes; Sentiment analysis; Social media; SolVES