Diversity of places and people: Using big data to understand languages and activities across geographical space

Urban populations are becoming highly diverse, or ``super-diverse'', through increasing globalization and international mobility. Super-diversity implies diversity that occurs across multiple variables such as language, ethnicity, religion, gender, age, country of origin, mobility, employment, and housing career. Language provides a useful, but underexplored perspective to super-diversity, as languages mediate every social interaction in urban areas and constitute a central part of individual and group identity. Superdiverse populations increase the diversity in urban areas also through their activities related to leisure, work, and everyday errands, all of which also vary across geographical space and time. I use the term urban diversity, by which I mean the diversity which emerges from the presence of super-diverse populations and their activities. Urban diversity exhibits spatio-temporal variation due to people's everyday mobility and the change of their residential areas. In this thesis, I concentrate two variables of urban diversity: languages and activities. Social media and population registers capture information about urban diversity, such as languages and activities, to different degrees. To better understand spatio-temporal urban diversity, the use of several sources of data and interdisciplinary approaches are necessary, as this understanding enables urban planners and decision-makers to support social cohesion and social sustainability in our cities.

In this thesis, I explore urban diversity from the perspectives of languages and activities using social media and population register data, focusing on Finland and especially the Helsinki Metropolitan Area (HMA). As urban diversity is a complex phenomenon, my work draws conceptually and methodologically on several fields of research: geographic information science, urban geography, and research on urban multilingualism, which also covers linguistic landscapes research. This thesis is strongly methodological in nature. I use diversity metrics originally developed in the fields of ecology and information science to assess the diversity of languages in population registers and social media content across Finland and the Helsinki Metropolitan Area. I apply computer vision techniques to extract information on activities from visual social media content. Finally, I use techniques from spatial analysis and statistics to examine the spatio-temporality of urban diversity across geographical scales from the national to local level.

I report the results of my research results in four articles. Article I explores the spatio-temporal diversity and richness of languages used by Finnish Twitter users from regional and user-based perspectives. The article shows how language use and linguistic diversity on Finnish Twitter varies across Finland, and characterizes the diversity of the linguistic repertoires of the users. Article II shows how to extract information on activities and visual preferences with several computer vision techniques from Flickr photographs taken in Finnish national parks. The focus on visual social media content circumvents challenges arising from multilingual and textually limited content. The article shows how the activities and landscape preferences of domestic and international visitors in the parks differ across the parks. Article III examines the variation in linguistic diversity in the Helsinki Metropolitan Area from population registers and social media during 2015. The article demonstrates how linguistic diversity derived from first language information in the population register and the linguistic repertoires of social media users can be used to assess where encountering a language other than one's own is likely. The article also explores what the background characteristics influencing linguistic diversity are. Article IV examines the spatio-temporal patterns of linguistic diversity in residential areas of the HMA between 1987–2019 and the integration of two sizeable local minority groups, Somali and Estonian speakers, to the Finnish society from the perspective of languages. The article reveals that while linguistic diversity is rising in all neighbourhoods, speakers of Somali and Estonian are exposed to it differently. The article also shows how linguistic diversity has changed in terms of the languages that constitute it and the locations where it is concentrated. Finally, article IV demonstrates that linguistic diversity in moderately diverse neighbourhoods is more likely to change, whereas monolingual and multilingual neighbourhoods are highly likely to remain as they are.

My results show that urban diversity is a spatio-temporal phenomenon and language is a useful variable for bringing out spatio-temporal patterns in urban diversity. My data shows the HMA has diversified rapidly from a monolingual area to a multilingual one. Moreover, the languages spoken in the HMA and the locations of multilingual neighbourhoods have changed, and changes in linguistic diversity in residential areas are influenced by their spatial surroundings. Furthermore, my results show that social media data reveals a more diverse spatio-temporal linguistic view of the HMA compared to what population registers demonstrate. Such a dynamic view provides more understanding of where and when urban populations encounter diversity. These results emphasize the importance of understanding the emerging spatio-temporal and social patterns of urban diversity, which provide vital information for policies fighting segregation, social tensions and social polarization. My work demonstrates the value of combining several sources of data, analysing them using interdisciplinary methods, while drawing conceptually on several fields of study to better understand urban diversity. As urbanization continues globally, and is accelerated by the climate crisis and increasing global instability, it draws more people into cities, interdisciplinary approaches to examining diversity in urban areas have become necessary for supporting inclusive, socially sustainable, and resilient urban futures.