The pleasures of all forms of mapping respond to the deeply satisfying ends to which they process ties across space for viewers, allowing them ways to access space through spatial networks that one never before had the pleasure to observe: and watching these connections and webs appear in a graphic template gives one a new sense of imagining travel across space, but also an almost innate pleasure of seeing how places are tied. The multiple frogleaps between shores and across countries in the above world map suggests a shifting notion of the inhabited world. With Facebook at its tenth anniversary, the immense amount of data that the social network generates offers opportunities to map impressive patterns of interconnectedness which shift attention from those borders which have long defined nation-states or indeed the bounds between land and sea.
In ways that might well be signs of the times, and surely aim to chart a sort of post-modernity, the maps of connected identities offer striking ways to see the world less in terms of its divisions than the links web-based communication allows one to harvest. Indeed, without depending on the nation as the unit of meaning: Facebook’s project of comprehensive social graphing has maps populations and social networking by tracking a combination of individuals and aggregates over space–at least, across self-reported locations. Facebook has already showcased its data as a means to map friendships in ways that suggest a new way of reporting linkages in the very sort of social networks it promotes. If all maps map human knowledge, these data maps also visualize space outside of spatial constraints: but a new type of mapping emerges by tracing populations on the basis of how its users self-identify as birth cities and cities of current residence. For Facebook engineers map actual spatial distributions by big data of the sort economists admire measuring, but most have found hard to discern. They are worth attending to in some detail, not only to ask about the data on which they are based, but to examine the beauty of their construction. By charting patters so compelling, they also force one to wonder whether the real story they tells go beyond the reach of the world-wide web–and what ways to best understand the geographical connectedness they actually reveal. Perhaps we can begin by asking what sort of vertiginous views they offer of the world and its inhabitants, and how we can best relate to the stunning images that Facebook engineers seem to have taken such pleasure to produce–and whose visual economy seems so pleasurable to view.
After a decade of storing Facebook-generated data in a thick dossier of computer files, Facebook claims the ability for a sort of aggregate mapping, but one that erases individual stories: its aggregates conceal the somewhat obvious fact that it tracks the self-selected members of Facebook users, and showcases them as if they offered a comprehensive enough record of the inhabited world to base a definitive social graph of global populations. Perhaps Facebook users are the population that we want to notice, or the population that it makes sense to track–but this limitation of data is a liability as much as a strength–and because it treats the usual boundaries that have long offered practical constraints to travel and exchange as superseded, aside from the unstated boundary lines that divide China or North Korea from the other countries of the East. The map suggests a new sort of human knowledge of connectivity and of the different sorts of markets, presumably, that those who are web-connected participate and are attracted by or move within. The map that results is striking both because of its erasures of the border-lines and boundaries–and maps webs and arcs of population links without clear reference to spatial directionality. The skein of ties that emerges has become somewhat iconic, and has also provided a sort of template or inspiration for the map of users’ “migrations” or shifting geographic locations in ways considered below.
Given the intentional elision of individual story-lines that is characteristic of the data being harvested, it’s important to try to untangle the maps it makes from the data on which they are based–if only to understand the bias of the picture they present: a demographic restricted to their users, as well as based on data that is self-reported and without any objective control. Rather than tracing the unidirectional movement associated with moments of migration driven by famine, slavery, war or ethnic tensions, the strikingly geometric maps Facebook engineers created reveal patterns of connectedness of special interest in tracking geographic mobility, but also in mapping (and perhaps describing) the new patterns of interconnectedness Facebook engineers would probably like to associate with the Facebook networking site. The oddness of using the self-generated information on a site of social networking as demographic data may reveal a thirst for processing big data; once invested with the aura of objectivity, the collective patterns it reveals can erase individual stories as much as synthesize completely credible and coherent narratives.Connectedness is fascinating to track. And such information like place of birth and city of residence are not considered categories massaged for personal advancement. But the internet offers enough screens of concealment and opportunities for disguise that the social graphs Facebook makes tend not only to lack objectivity, but base themselves on data pontentially massaged by users–who, living in the suburbs of one city or in a nearby village, may report themselves as living in the best-known largest nearby city readily recognized, less in an act of deception, but in ways mistakenly treated as objective in character.
If all maps tell stories in how they embody meaning, the stories that the maps based on Facebook data suggests demand unpacking, if only since their numbers are so apparently comprehensive and large, as well as the stories that they submerge. One story they promote, which we might not accept, promotes an increasingly inter-connected world; others that it conceals is the specific stories about who is connected to whom and why. With almost half of Facebook users identifying two hundred friends, the webs created by the data purely of Facebook connections reveals an inter-connectivity that privileges the densest distribution of populations in North America and a broad sense of Europe, with a particularly active groups of users also dispersed in Central America, Peru, Argentina, Anatolia (Turkey), Nigeria, South Africa, Pakistan and India, and Indonesia. This map of global “friendships,” developed by Facebook intern Paul Butler in 2011, offers a “social graph” of some 500 million people, linking users’ friendships to actual coordinates of latitude and longitude to create a visualization in a sort of media mash-up, illuminated by an intense if other-worldly backlit glow of FB blue. The map of social networking that results suggests not only who is networked, rather than density of habitation but of course also maps the density of the culture of social networking that has come to stand, especially for Facebook, as a sort of index of modernization. The equality between Facebook use and modernity recalls how closely the ‘Twitterverse’ maps onto the maps onto the global flight-paths of airline routes, as new study has found, suggesting that airline connections between cities provide a statistically valid proxy for twittering to sites beyond one’s own hometown–and the ties that Twitter makes in virtual terms should not be in any way that distinct, after all, since “Twitter is part of the real world” as is the internet, as Matthew Battles astutely observed, and tweeting is shaped by real-world social ties, even if it offers a new media for communication. But it would seem a misleading inversion of cause and effect to take tweeting as an out-and-out proxy for modernization, equating it with the tracing of ties commerce and capital flows.
The question is not about the data of Twitter vs. Facebook, but about how big data is used to create the multitude of variant maps generated by Facebook media, including this map of ties among its users. The multitude of variant maps generated by Facebook media seem particularly striking for their post-modernity in how they elide or ignore the frontier as a primary unit of meaning in world maps. Is this map an illustration of of the acceptance of Facebook as a part of one’s identity, and do the great arcs themselves constitute something of an accurate index or emblem of post-modernity? While Facebook is actively interested in promoting is own ability for harvesting data, many of the maps of friendship, migration, and social ties of course read a lot like images of its ability to promote social networking. The arcs of light that map online ties not only mirror the sort of etherial airline map that seems as good a predictor of twitter ties–does this suggest the similar likelihood of access to airplanes as access to tweeting?–but a gauzy penumbra spun around the earth’s inhabitants who sit spellbound before backlit screens, luxuriating in the proxy embrace of on-line connections and affections? It’s always important to ask what sorts of abstractions that any map seeks to communicate through the tenuous bonds they draw linking actual regions.
Is this a new story the habitation of the select areas of the information-sharing world that Facebook links? The above image tells one story about the inhabitants of the world, or rather of a specific world, and a strikingly suggestive one in how Facebook’s data defines the worlds inhabitants as its users. For there is no doubt a certain amount of estrangement from other forms of communication that this image, linking pairs of cities weighted for how many friendships exist between them, also maps, varying the intensity and brightness of lines’ coloring by the number of friendships shared as a way of mapping “real human relationships” across space in ways that Facebook does not note explicitly: the map devised by Facebook’s infrastructural engineering team is a map of the very distances that Facebook as a medium is best able to bridge. If perhaps just as revealing of access to computers, digital literacy, and wealth, the arc drawn across the empty spaces of deep blue that are oceans charts the networks of inter-connectivity among which news travels, memes grow, personalized photos shared, and a sort of quasi-information society exists, by the construction of the social network on which it travels. This “globology” reflects less the currents of cultural contact, migration, and exchange that define global history, but something analogous to the cartographical flattening of it, viewed through access to the options of like, connect, dislike, or share that Facebook offers its users. It charts the options that have shaped, to one extent or another, a transnational psychology of communication.
Which makes the ambitions of Facebook engineers to map patterns of “coordinated migration” an interesting case in mining of new trends in geographic mobility from the massive amounts of data that they collect, if only because of its goals to map actual spatial position over time. The somewhat-recent maps of what FB term “coordinated migration,” based on the patterns of relocation that emerge from self-reported data of the site of birth and current location of FB users. The result of the database’s collation is to track and define the lines of the global phenomenon of urban migration, once reserved for larger industrial cities or metropoles, but not significantly broadened across the world, both to population growth and a drying up rural prospects and jobs in the magnets of urban metropoles. The idealized format of these arcs of migration create a social scientific aura of explanatory simplicity, and serve to affirm the notion that individual itineraries respond to an economic marketplace. The boast of Facebook to chart such pathways with accuracy has been critiqued for its use of private data without evaluation through the scrutiny by peer review, and may deviate from standards of impartially objectively derived data, or by using publicly listed self-identification to endow with the graphic objectivity (or appearance of it) on a map. The maps make tacit claim to reveal the real attraction of populations both from neighboring cities in the same country, or, in South East Asia and India, from neighboring countries, by the web of tiny green arcs discussed in greater detail below, provide a useful filter for sort of big data economists have problems measuring of global migration, and indeed to distinguish quantifiable migration numbers and patterns in different regions of the world. Once again, frontiers are less significant than the ties of cities that the data input by Facebook users enter into their pages, and which they map.
Underlying questions remain, but seem to be waiting for further or future analysis. For one, how, Sheryl Sandberg, do these patterns of migration map out along gender lines? In mapping “destination cities” by red dots, and “origin cities” by blue, the design group Stamen has helped render a smooth arcs of migration as a set of naturalized movements, in order to throw into relief the relative density of a phenomenon of “coordinated migration” around the world, but creates a naturalized web. They give one useful key to read meaning from these idealized arcs by setting them against a global map coloring regions of relative urban growth in 2000-12 pale yellow, in the sort of compelling image compressing a density of variables to simple graphic forms of which Tufte would be proud: routes replace cities that they link, or that appear as numinous purple blurs that remind us of the messiness of mobility; the crossing of frontiers replaces frontiers, which must be deciphered by the limited role of crossings from China, whose very existence is only implied by paths of emigrants abroad.
The mottled blotches of blotting-paper blue to which the thickest green trails lead indicate the fastest growing markets of production that provide the magnets from nearby countries as well as for regional residents. The linking that the map tracks suggests a sense of populations in flux, and of the market-driven nature of the migrations that Facebook’s data tracks, and whose intense mobility is a sign of relative modernization as much as social unrest. The broader linkages such data implies between Facebook and the global process of modernization is, certainly, good publicity, no doubt, if largely on the level of a subconscious association.
But the emergence of what Facebook has christened in quasi-sociological jargon as “destination cities” might be taken with a grain of salt. But the data that Facebook has so far collected is also an inversion of a somewhat embarrassing incident that occurred at Facebook back in 2010, when Pete Warden devised with the inventive project to scrape the public Facebook profiles of 210 million users in the United States to derive the clustering of networks that divided the country according to FB ties, creating his own national map, but then withdrawing the offering of the data as public when Facebook threatened suit, but had to allow him to keep up the map he derived from the data–brilliant if only for its definition of clusters that would never be rendered in a printed map, with eponymous if imagined toponyms as Socialistan, Mormonia or winning ones like Stayathomia, designating the greater northeast and midwest–a region viewed by most as revealing less geographic mobility, but which wouldn’t appreciate that name.
The map, soon taken down, may have encouraged a sort of counter-posting of a heat-map that announced the world-wide distribution of Facebook developers, which seems designed to celebrate their global geographic diversity, rather than social or geographic segregation, to dispel the notion that they are concentrated in the United States or, in terms of the city where most FB developers live, San Francisco. (The coloration of whole countries based on local concentrations of software developers finds all of North America painted red.)
But the huge amount of data that Facebook has developed in a decade-long harvest is hard not to want to map, if only to see what sort of visualizations it produces. And the three or four maps they issued tracking the benefits of ten years of social networking have led to the claim to reveal new patterns of social migration to cities or urban metropoles that are the equivalents of hot-spots in a globalized world. The shading yellow of regions the World Bank found marked by high “urbanization growth” over the same period–2000 to 2012–creates a simple context to situate the migration, although it would only really make sense to chart that growth in population over a broader period corresponding pressures that impacted those who were migrating in ways tracked during this period as a way of registering, or so Facebook would like us to think, some of the primary forces that have reconfigured a globalized world on which Facebook has a clear take. The cities noted as cross-national destinations reveals a sense of global mobility and inter-connectedness with which Facebook wants to promote, to be sure, linking a sense of globalism with its product’s global reach, but wrap the world in a skein of green lines link with differing intensity selective cities an towns, suddenly giving prominence to the links of a potentially economically prominent demographic.
The elegantly arcs of migration charted from original residences are quite impressive localized illustrations of geographical mobility: if the result might not be truly comparable to how Aude Hofleitner, Ta Virot Chiraphadhanakul and Bogdan State have decided to evoke the many “large-scale migrations [that] are an important part of human history,” the chart reveals micro-patterns of geographic mobility by neat green lines would make a social historian turn green with envy, mapping “flows” from hometowns to current cities. The responsible social historian would ask whether those nice curved swirls map along a symmetry misleading in portraying actual paths that those migrations took–many, one could speculate, were routed through multiple other cities or post-modern metropoles, as the line connecting any two points anywhere is rarely a perfectly smooth trajectory or arc: the visualization betrays, in this sense, the data-centered models from which they derive, and are, in a sense, less revealing of a picture of actual populations. But the floral patterns they create, spinning from blue-dots of geographic origins to red dots of final destination or current habitation, document a fairly radical shift toward urban in-migration, reminiscent of nineteenth-century Europe, and a huge attraction to sites in coastal west Africa, not only Lagos, whose urban population’s growth by almost one-fifth reflecting the expansion of the city and oil-production, but also Dakar, Accra–whose urban growth also approached 20%–, Ghana, the Gulf of Guinea, and Nairobi, as well as Dar es Salaam. (One suspects that this is but a shadow of the true migration of populations, since it only represents data from the social network that Facebook runs, and is silent about its very selective picture of folks with internet access and net-literacy.) But the migration of Nigerians it tracks to Lagos is nonetheless prominent.
Could this data not be usefully cross-mapped with the growth or social compositions of destination cities in more provocative ways? What sort of visualization would that create? Can it be refined, by accessing data like the changed place of Facebook users over time, creating a somewhat more messy, but perhaps more informative, way of tracing patterns of migration, than a graphic that seeks to underscore as well as illustrate the notion of a “coordinated migration” from smaller towns to metropoles? To be sure, the arcs used to denote the alleged concentration of migration area a way of softening the image of Facebook’s data, making it more organic part of the map, and not serve as an apparatus of any accuracy. The arcs are aesthetically engaging as a way to illustrate each “destination city” by its reach to its hinterland–or to other cities of origin–to visualize and showcase the massive amounts of data that it has collected, much in the way that Facebook has engaged in multiple ways to visualize the scope and extent of data that it generates over a huge expanse of the world, if not to suggest it forms a way of tracking the inhabited world, under the name of Data Science, always keeping that data anonymous.
The map of major destinations of folks in the two year period of 2010-12 is striking for how it reveals a changing face of global socio-professional mobility that seems specific to some classes. The usage of circle arcs in the streamlined OSM cartographical data is almost as misleading as it gets if one thinks that the arcs mirror the messiness of routes of actual spatial migration. But the abstraction of defined population currents offers an abstraction of presumably “eminently hirable” individuals looking for more profitable lines of work, linking the Facebook populations with the up and coming professionals of the future, as much as with job mobility. The yellow highlighted areas of notable urban population growth create a backdrop that increased the legibility of this slightly restricted map, and others below–one wishes for a bit deeper patterns of urbanization and rural pressures–at times far preceding the decade we are describing in these maps.
The graphics are particularly striking visually, and effective as compelling images of global patterns, even if they also might not be that deeply informative: they trace a collective social history of urban in-migration in Istanbul, the city Facebook’s researchers found to have the greatest “coordinated migration,” largely as it is the greatest metropole in a largely rural but rapidly “westernizing” country whose infrastructure has greatly transformed–as if Facebook use was a predictable index of economic modernity and, by extension, growing markets for its advertisers. The foci Facebook’s data has illuminated here indicates as “growth towns” such destinations as Nairobi, Lagos, Port Harcourt, Accra, Kumasi, Kampala, and Dar es Salaam, but purged of the living conditions in those sites.
But a more complex super-imposition or layering of not only urbanization, but annual rates of populations’ change worldwide, in ways that the United Nations has also charted in ways that would be of particular importance for such coordinated migrations, which focus, say, not only on Mali’s and Kenya’s urbanization, but growth in Zambia, the Central African Republic, Chad and Tanzania–and explain why these are to differing degrees loci of the very migratory pressures that Facebook has detected. (All data deserves to be mapped, but the Facebook seem to privilege almost exclusively the very data that they generated, in order to showcase the trends that its use is able to detect, but some maps are misleading enough that they are perhaps less better drawn. This leads to some complications, both since the data is not made public and controlled, but also since it is not able to be combined by other data or metrics in responsible ways.)
The transformation that these maps chart might be one of the huge expansion of internet users–the maps map their own data, after all. The clustering of such users in a rapidly growing city like Istanbul is particularly compelling, and interesting reveals the huge amount of in-migration from the less sparsely populated areas of Anatolia around the Black Sea that led to the city’s growth by almost 12% in 2010-12. It would especially profit from a charting to areas of the city–if possible–or the expansion of the city itself, as well as the growth of other Turkish cities from which those arriving in Istanbul hail: are these towns declining, and just dying, or are they also growing at a comparable rate? Do folks moving to Istanbul list actual home-towns, or the nearest large villages? Istanbul is perhaps also unique in being a modern metropole drawing educated populations from many nearby Macedonia, Bosnia, Bulgaria, Albania and Hungary, as well, it seems, the Ukraine–in a sense of the rebirth of the Ottoman Empire, but in ways distinctively lopsided to the East: a blanket of green gauze of strings of differing intensity whose arcs blankets the northern regions of Anatolia bordering on the Black Sea.
The simplified elegant light green arcs in the above map and its above counterparts stand in quite sharp contrast to the rapid circulation of populations mapped in what is by far the densest and in senses most complex map groups India, South-East Asia (Thailand, Cambodia, Vietnam, Philippines, Singapore, and Malaysia)–almost a data-overload itself–or of the specific stories of these immigrants. Of course, the question of why this region–embracing really at least three, but probably four sub-regions, should be seen as a collective is not clear, but it is essentially “the rest of the world” aside from the main regions Facebook has parsed. The graphic suggests a far more massive amount of mobility that is almost difficult to distill to a single image or map. (China is absent from the dataset, because internet use is so less common; public self-identification is also no doubt far more rare, and foreign, anyways.) There was quite explosive growth in this period in Bangkok (significantly over 10%), Hyderabad (a bit under 15%, or 14.4%), and Chennai (also slightly under 15%, or 14.4%); one wonders more about the social demographics that motivated this growth, and how much of it maps onto the Facebook users who it maps.
The loss of the stories of individuals who are immigrating to cities is almost part of the point of these maps: there is no story that needs to be told about this immigration, because it is a process of migration that responds to uniform global pressures we all recognize, which are specific to the demographic of Facebook’s users. This sense of a story is far from Satyajit Rays’s vision of “The Big City” of the immigration of rural villagers to their new life of searching for commercial employment in the banks, mechanical sales, and lawyers electrified Calcutta. For the migration to the city that dominates each regional state–rather, the data charts the growth of emerging metropoles as Chennai, in southern India: it charts mobility from cities to cities, inter-city mobility, among a I-think-that-I’m-probably-quite-employable demographic of relatively educated Facebook users who are most often online, and would coincidentally provide a selective demographic for advertisers to reach. Indeed, large swaths of rural territories are essentially unmapped, save by the intersecting bright green arcs that cross over their terrain, as if without Facebook users, they did not deserve to be mapped–or, at least, without necessary data, fall out of it–as if to locate the intensity of migrations around its rim, although quite a few sole threads of migration originate from its empty interior.
The image of a growing South Asia is, however, particularly interesting for the number of itineraries it traces out of China, migrating to Taiwan and Hong Kong for the most part, as well as South Korea and the Philippines; the active clustering of Facebook users along the coast suggests that it is were the true action is occurs, as if to rewrite the attention usually paid to charting China’s escalating economic growth. The vibrancy of ties across borders (and across seas) is striking, as a certain demographic seems irresistibly drawn by Taiwan, and others to Thailand (Bangkok) or Cambodia and Vietnam, whose arcs reveal notably dense ties of geographic mobility.
Whose stories is revealed in this amazingly dense map of neat green arcs, which, if they simplify the messiness of the individual stories they collectively track, seem intended to create a sense of a coherent network of meaning that might not be present in the area’s economy but is actually extra-national in nature: these migrations seem links of like to like, and of educated to industry. One however wishes for greater resolution in an image such as the above: what is going on in the Philippines, for example, where Manila is not so much the central hub one might expect, but that each island seems to have its own centers of aggregation and congregation, is unclear; out- and in-migration also seems mixed, as it does around Djakarta and to an extent in southern India and Sri Lanka. Would a close up, say, of Polynesia or at least of the area around Singapore be of help in discerning what patterns of “coordinated migration” might emerge? The huge urbanization of Java, Bali, Indonesia, Laos and Vietnam, and the circulation in populations in the rapidly growing region, as well as any other political and economic pressures on this sea of population shifts. Many of the arrivals in Bangkok are from within Thailand, moreover, rather than from other countries, suggesting a very traditional pattern of migration within linguistic and cultural units, rather than the global circulation that such maps might be expected to chart, or might at first glance appear; similar interior migrations characterize Vietnam and, it seems, the Philippines, where much migration even appears specific to each island. How, to use a metric that Facebook has cleverly adopted in its map of friendships across nations, can one view these ties through the lens of shared languages or linguistic similarities?
There appears considerably less defined patterns of migration, it bears observing, in the United States. What does this mean, or might it indicate? (Or, which seems likely, are Facebook users in the United States just less interested in noting the cities where they live after they geographically relocate, by updating their profiles?)
The patterns provide a new way of reading ties within the map that trump, or erase, geographic proximity, in ways that seem of interest in tracking, for example, the shifts of economical mobility that might be tied or associated with NAFTA. What is one to make of the far lesser webs of connection or migration in the North America, where there seems, save for the Cubans who have arrived in Miami, far less social mobility or a zero-sum game? Facebook works–and thinks–in coordinated networks in this case, and finds little similar coordinated migration within North American cities; even in the international migrations from Mexico and Cuba, which seem to be potentially overstated–they reflect the number of Cubans in Miami that state their origin as Havana or Cuba; one can detect less of a coordinated action than a mass movement, in the case of migration from Mexico, than what might be of a phenomenon chain migration, rather than one independently coordinated, and determined more by family links than economic need. Three different patterns of migration seem to be evident in the three countries shown below–even though the Stamen projection seems to erase national boundaries or borders by rendering them as equivalent to regions or states: Canadian cities draw amost exclusively from the surrounding province, unlike most of the US, save North Dakota and New Mexico; Mexico shows active cross-national migration patterns, perhaps based on familial ties, the engineers note, to such major destination cities as Chicago, Houston, Dallas, and Los Angeles.
So do the above mapped elegantly detailed distributions, given the selective nature of their self-reported data, mean anything definite at all? The skein of green lines of differing intensities wrap the world in a variety of ways that contrast, for a start, with a simple map of flight paths: what remains untold, however, is exactly what demographic the Facebook family has been able to attract.
It is hard to say what they mean exactly, as a result, or how much they provide indices of mass migrations, even though this quite valuable data should by no means be dismissed. Since it depends on self-reporting of information, they capture a small number of the actual migrations that occur between countries, by migrants who are less likely to log their travels in internet profiles or lists of friends, but the data that the engine of Facebook has, ten years into its existence, managed to compile, and is now eager to show the world that they are able to track. Of course, the data is only data–and it’s of interest to create a map of data, if only to see what results. And the somewhat over-determined data of who uses Facebook–folks who are geographically removed from their loved ones?–may be difficult to map as a trend. It provides a compelling collective visualization among folks who belong to that demographic that often reveals the persistent meaning of the local and regional in a so-called globalized world. Perhaps it doesn’t really map a world that is less rooted in geography–Google, after all, generates the GDP of a small country in itself, after all–and where place isn’t supposed to be so great a determining factor or constraint in most markets and lines of work. The question is what such trends of migration–if that is what’s being charted–mean, and what corner they raise on a map of larger trends of relocation.
Musings on Maps 