During the past months I’ve been working with urban planners and policy makers in order to identify the most meaningful urban questions we can match with the possibility given from the interpretation of those data.
First conclusions enligthen a few main area of interest in:
- Political Attitudes: acceptance, feeling toward local policies and urban interventions: understanding how new urban policies and infrastructures are evaluated and being able to plan new public services for specific groups and places;
- Behavioral Mechanisms: discovering possible emergent programs, structures and bottom-up initiatives responding to uncovered needs and desires as well as predicting gentrification areas and patterns of use of the city;
- Places identities and relationships: delimitation of territories, links between areas, emergent nets and relationships and connections between places, people and uses;
- Perceived well-being/Livability: alyzing users’ perceptions related to specific geographic areas and providing attractivity and urban qualitative indicators to compare diverse zones;
- Temporary events: understanding how people, distributions, emotions and uses changes over the time, relating contributions to specific events, periods or situations.
Furthermore, due to the profile of the people that actually use Social Networks, the most promising domains of applications could be summed as follows:
- tourism (temporary inhabitants profile, patterns of mobility, needs and desires);
- planning the cultural offer of the city (which places are named together with ontologies related to culture, emergent local and global net of places related to cultural activities, cultural influencer);
- temporary citizens (emergent ethnic groups: who they are, where they are, how they use the city,)
- the rhythm of the city (time based analysis of the temporary cities: how does fluxes, areas, concentrations, profiles through time?)
- public health (identifing qualitative indicators able to reveal people perception about specific areas)
- mobility patterns (identifying fluxes and pattern of mobility of certain groups of people within metropolitan and urban areas);
UGC differs from conventionally produced geographic information in several aspects: the source of the information, the technologies for acquiring it, the methods and techniques for working with it and the social processes that mediate its creation and impact. Traditionally, geographic information has been produced by experts and institutions: so certain types of information have been privileged and other types ignored or marginalized. UGCs represent a powerful shift in sources, content, characteristics, modes of data production, mining, sharing, dissemination and use.
Therefore, a wide set of research questions (that have been partly investigated for “conventional” geographical information) need now to be re-investigated. Even though UGC are not representative of the whole population they are an actual trace that has to be investigated:
such novel sources of information need then to be compared and overlapped with traditional ones (like ethnographic survey and on-field observation, data provided by census, by the public transport or by telephonic companies) or interpreted for their own specificity.
The data we gather need then to be analysed starting from specific urban questions: mathematical models as well as statistical indicators need to be developed for each precise case study.
It is possible then to build a set of indicators (knowledge patterns) to be used to derive useful knowledge?
I present as follows an in-depth examination of the area of interest we discovered.
(*these conclusions arise from a work that has been conducted for the writing of 2 different proposal we submitted to the European Commission)
Political Attitudes: acceptance / feeling toward local policies and urban interventions.
Through an analysis of UGC, it would be possible to understand how the population (social network users) react to new policies. The ways in which these reactions can be expressed can be categorized within a concept of customer satisfaction. Two primary indicators can be used: one determining the “attitude” (e.g., Is the feeling toward the introduction of a new pedestrian area positive or negative? are people happy with the new bike lane?) as well as one which attempts to gauge a more interpretative contents of text on a discursive level e.g., How are people talking about a new proposed policy intervention, what terms are associated with the public interest issues such as quality of the public transport service? Do they agree on the evidence produced by the institutional decision makers about the need of the new intervention? Do they conversely prefer a different approach to the problem? If so, which? Moreover, does the whole of the popoulation react the same? Can we enlighten differences among different ethnic groups, social groups, profiles and gender, ages, or from people who inhabits different part of the city?
Behavioral Mechanisms: emergent programs.
Very often people are not offered what they need from institutions and administrations. This leads to emergent structures, bottom-up initiatives meant to answer eluded questions, needs and desires.
Another component which influences behaviour as expressed can be measured in the level of communication and degree of information sharing between government and citizens (e.g., what are the responses to public officers’ tweets, Facebook posts etc., are new ways of public responses emerging – from a traditional public protest and campaigning to expressing opinions in digital formats) These two thematic layers refer to the most explicit side of the local knowledge that can be recognized in the UGC analysis, since it refers generally to the rational discourses and behaviours of the subjects involved in the analysis. Is it possible to discover the lack of infrastructures od a city by analysing users’ on-line complaints? Apparently in Milan there are no mosques. Obviously this is not true: where do people talk about religion? Which places are named together with the name of God?
Moreover, the nature of the data sharing (profile of people that share) is very promising in order to intercept and predict gentrification areas, as well emergent places and people able to drive trends at the local and the global scale.
- Places identities and relationships.
Beyond the most explicit and rational part of peoples’ beliefs, convictions and desires, there is a more intangible part of knowledge that has to be identified by exploring different dimensions of urban life. One of these is the ties to the urban communities. In every European city, there are areas with strong character and identity. It’s clear to everyone where and what a historical centre is, how to define an urban park or a hilly countryside. But how can we define and how people name areas in transition? Which are the most named places? Can we identify novel categories for interpreting our cities? …novel urban landmarks, districts, nodes, path and categories… Which places are never named and why? Where people recognize a territory as peripherical? Which are the boundaries of the actual “center” of the city, as named by people?
As multi-locality, international and transnational as well as “cosmopolitan” identities are being expressed among citizens, often more than one at a time, it brings into to mind the notions of who is part of a global identity and what is it comprised of. For instance, the ways in which people represent themselves in reacting to threats or disasters, issues such as migrant integration or social justice, economic immobility, constructions of “others” or identifications of “sameness” can be traced and interconnected. The transnational identity of mobile youth through university, workers both in and out of their countries of origin, for instance, and their representations in the local media, neighbourhoods and art have created both social and new policy challenges.
Furthermore, how can we intercept emergent links between specific areas and places? Those nets defined by the pattern of mobility of certanin groups od citizens that select specific places. UGC has the potential to be used as an alternative means of gaining knowledge of shared concepts of space.
Perceived well-being/Livability: attractivity.
What is that make people talk about an area? Shopping? Theatres? Its food scene? How many different activities make people talk about an area? How many thematizations are possible in a district? How do things change over time (on a daily, weekly, monthly and yearly base)? In other words, what are the public’s perceptions of a few central elements often associated with “good” government and thus high livability of a city. This could be done with an initial scoping or “digital emersion” phase where the research can identify the appropriate mechanisms, places and ways to build relationships based on their cognitively mapped, physically mapped, and socially mapped networks of social hubs and focal points where further participation could be elicited. These links should be traced in order to see where a social change agent (or blocker) has been identified through an expression of values of what livability in certain geographic areas means and what is needed as expressed by the inhabitants within the community, as issues, alternatives, and their position on the current status of the livability. Through the application of sentiment analysis techniques within the textual analysis of UGC, it will be possible to identify generic sentiments from the citizens toward their cities, specific neighbourhoods within the cities or single urban infrastructures and urban projects. Sentiment analysis will allow to extract information about, on the one hand, users’ judgment and evaluation, classified according to their polarity (positive, negative, or neutral) and, on the other hand, about the affective/emotional state of the writing author, e.g. happiness, fear, etc. While there is extensive research about sentiment analysis per se, and a growing interest for applying sentiment analysis to UGC and social networks, the application of sentiment analysis to spatial data has not been investigated yet.
- Temporary events:
Physical spaces become places as they are experienced, lived, shared and communally interpreted. Contemporary urban analyses are increasingly focused on micrological and interstitial investigations to investigate the temporary, precarious, and conflicting meanings attributed by specific groups to places with which we often maintain a familiar relationship: a neighborhood in our city, a park, an event, a series of paths. These places become ‘practiced places’, to quote de Certeau, or in other words spaces interpreted and lived through the experiences of situated subjects (de Certeau, 1984). Thus, bringing a temporal dimension into those data systems gives us the possibility to compare spatial distribution peaks as well as emotional peaks and recognized patterns to specific events. Simply analyzing the spatial and temporal distribution of different groups of people we can answer to meaningful questions such as: Where are people from? How many tourists ara talking about the event? how different profiles distribute themselves within the city during a spread event? From where people talk about the event the most? Moreover, with keyword based analysis we can intercept how (when, and from where) people talk about the event even before and after it. Which brands / commercial places / people / entities / organization are named the most? (we’re currently experimenting on the MilanDesignWeek and we’re monitoring what’s happening online with the issue of Expo2015.)
*How can this data be validated, in order to verify if they are significant and accurate proxies of common perceptions toward city spaces?
*How can well-being and happiness be defined and observed from user generated content? And what use can we make of ubiquitously connected things and sensor data of various kinds towards the aims of the stakeholders?
* How can be the UGC used in depicting the two most important components of the localized knowledge: the ‘know how’ and the ‘know why’?
* Which knowledge representation structure (e.g. ontology, decision tree, etc) is required to map all the different sources of information and knowledge about the urban domain?
* Is it possible to plot the very many and co-existing perceptions in cities and neighborhoods, relying on UGC-derived knowledge?
* How to integrate such data with other source of information from traditional media, from direct observations and questionnaires, from the municipalities statistical offices?
As a side result, which me consider not less of importance, using the results of such analysis to sensitize decision makers about the actual perception of the city and to drive citizens to a more participative and collaborative attitude seem a promising field.
Thus we’re exploring the many possibilities of returning data back to the people who shared them, closing the feedback loop form where data are originated to where they are showed.