A text-mining approach to assess impacts and benefits of Nature-Based Solutions

Worldwide, research community has studied the benefits of green and blue spaces implementation in urban areas, generating a great amount of literature regarding this topic. Since these solutions are of interest to face climate change impacts in cities, the European Commission (EC) has funded several projects to make an extensive review of the available literature. Three of these projects were especially studied here, namely EKLIPSE, Mapping Assessment of Ecosystem and their Services- Urban Ecosystem (MAES: Urban Ecosystems), and NATure-based URban innoVATION (NATURVATION). They all aim to identify the physical and social impacts, benefits and trade-offs of Nature-Based Solutions (NbS).

To objectively compare findings presented in the deliverable reports, a text-mining approach was carried out. This methodology coupled with a data visual representation allowed to convert the EC projects reports (corpus) into a meaningful structured analysis. As a result, a graphical representation was created, making possible to recognize concepts, patterns and attributes addressed by each text, as well as stakeholders and their position with respect to the topic.

 The text mining analysis was implemented through Gargantex Blue Jasmine Version (an open source software developed by ISC-PIF). Gargantex results permitted to recover a list of key-terms from each corpus based in their co-occurrence in the whole text. These terms were used to elaborate a visual representation or network, placing the words strongly related close to each other and characterizing the obtained clusters by a similar color.

This approach underlined the specific focus of each project: the conciliation between urbanisation and urban ecosystems (MAES), or the economic valuation and monetisation of NbS (NATURVATION) for instance. Moreover, it demonstrated that despite the different literature review methodologies of each report/project, there are some common trends exhibited by the obtained graphical networks and their statistical attributes. For instance, the need to assess the NbS performance with some adapted indicators; and the important EC supporting role in the implementation of NbS. Similarly, some regulating (e.g. water quality or temperature reduction) and cultural (e.g. recreation or health benefits) services are more addressed.  

This analysis can be applied to all kind of corpus, which makes it easy to understand different and similar concepts and approaches of a set of text data. A text-mining analysis can be conducted over the direct references of NbS benefits, on a collection of publications of a research database like Scopus or Science Direct.