Sunday 31 December 2017

GeoGuessr: Navigating real-world places, virtually

Have you ever imagined what it would be like to be transported somewhere in the world completely unknown to you? And how you would figure out where you are? In this post, I’m discussing a game that simulates this kind of experience virtually, using Google Streetview. 
GeoGuessr is a web browser game first developed in 2013, in which players are presented with a series of five Streetview images from five different locations across the world in turn. Based on what they see, players must guess the location that the images were taken by pinpointing on a zoom-able world map. After making their guess, they are assigned a score between 0 and 5000 based on the distance between the guessed location and the actual location the images derive from. The scores for all five rounds are then totalled at the end.
From a geographical perspective, GeoGuessr is thought-provoking for several reasons. The views of landscapes provided by the Streetview images give players the opportunity to see places they would not otherwise see, yet inevitably represent environments through – literally and figuratively – a certain kind of lens. Secondly, the act of finding and interpreting the information presented in the images is an exercise that reveals the individual processes and customs through which we make sense of what a place is like – the ways in which our relationships with environments are formed. And lastly, the design of the game itself works to turn the medium of Streetview images into an experience that – more or less successfully – enables players to engage meaningfully with different places across the world.
Let’s delve deeper into these three points of interest to unpick how and why GeoGuessr presents such valuable questions about what ‘places’ are in the digital age, and how we interact with them.
Exploring landscapes
As we consider the experience of ‘visiting’ locations in GeoGuessr and interpreting what you find, the key factor to recognise is that the environments you explore are communicated as images: moments in space and time, captured from a fixed point of view with specific equipment, with specific motivations.
Images are evocative tools for communicating information about places. So much so that, even if viewed a great distance away from the location in which the image was made, we can often still get a sense of what it is like to be in the place represented onscreen.
And it is their detachment from the moment in which they are captured that makes it possible for  GeoGuessr to expand the boundaries of exploration as a practice into the virtual realm. We are able to see places we would never physically visit in our lives, and get a sense of what it might be like to be there, at those locations. It is for this reason, in part, that GeoGuessr has been praised as an educational resource, providing learners with an extensive database of everyday scenes taken across a wide range of landscapes, and packaging it in gameplay that encourages you to pay attention to detail in your surroundings. With each guess, you can open your eyes to another small portion of the world that you are unlikely to ever visit in person.
Importantly, the Streetview images that GeoGuessr uses have their own distinct qualities. When viewed, the viewer can rotate the image nearly every direction from a fixed point, situated at roughly head height from the street. Additionally, the viewer can use onscreen arrows to navigate between the images taken as the Streetview vehicle moved through the site, as if you were travelling through the environment yourself. It is these two characteristics of the pictures that make the experience of viewing them more immersive than the fixed, linear perspective of traditional images. Unhindered by the borders of the image for both perspective and movement, Streetview pictures gives players a better appreciation for scale, and how nearby points in space are connected together. In short, the experience is much closer to how we would encounter an environment in the flesh.
Indeed, the creator of GeoGuessr, Anton Wallén, has said that the idea for the game arose from his enjoyment of visiting faraway locations on Streetview, and how the images could make the viewer feel as if they themselves are in the places photographed. On a more everyday level, many of you reading this will be familiar with the process of using Streetview to scout a future journey. For pedestrians and even those travelling by vehicle, the street-level perspective of Streetview’s images is often much more suitable for familiarising oneself with a route than the birds-eye view of a map.
And yet, navigating in Streetview is still a very simplified version of what we would experience by inhabiting a place ‘in real life’. Although we witness what is happening in a location through momentary scenes, the moving camera is largely detached from the realities of day-to-day life in the localities being photographed. As viewers, we are treated instead to approved snapshots taken at intervals along roads, taken during particular weather conditions, times of day, days of the week, and so on, which inevitably colour the mental model of the location that we develop in our brains as we navigate between the scenes.
And as we develop these mental models, we must not forget the context in which Streetview images are produced. Google’s cameras travel through the locations being photographed, and the images they capture are stitched together and uploaded to the Google Maps website for those with internet access to browse at their discretion. They are, first and foremost, images produced for the benefit of a multinational corporation and people living predominantly in the Global North, who more readily have access to the equipment needed to view them. That is to say, that the images do not represent a view of the landscape in accordance with the values of the community that lives there, but instead have been extracted for extraneous purposes.
In the same vein, there are many places that are ‘off the map’ because Google Streetview just hasn’t been there, for a whole host of political and logistical reasons.
So in spite of being a method of imag(in)ing places that is often construed to be more immersive and accurate than previous imaging/mapping techniques, Streetview ultimately reflects the same observations about landscape images that cultural geographers have made for decades. That is, that every instance of landscape representation is determined by relationships of power that influence the form used to develop the image, which places themselves are represented, the motivations behind these two details – and, in turn, how ‘knowledge’ about a location is produced.
Navigation and interpretation
The method of playing GeoGuessr involves a bit of detective work – using clues from the environments pictured, along with your own judgements, to build up an idea of where in the world the images have been captured.
Fortunately, because the game uses Streetview images, you can look in all directions and travel in any direction you wish (providing that the Streetview camera has travelled down the road in question). In nearly all cases, moving through the locality from the starting point is essential to get enough evidence to make an educated guess about where it could be. And it is this process of evidence-gathering, as you take note of significant details you find while moving through the stitched images, that you come to experience the intricacies which make any particular site unique.
It is a somewhat transformative ritual, because as well as learning something about the place being represented, we learn about ourselves – the array of memories, social norms, pre-cognitive experiences, and other sources of information through which we come to know a place.
That said, the characteristics of the information provided in Streetview (i.e. what can be perceived directly at street level) mean that certain kinds of details become more important than others for discovering your location. These are some of those most common types players use, which become familiar over time:

Road signs
Languages used on signs/buildings
Road markings
Terrain – vegetation, gradient, etc.
Cultural symbols – flags, icons etc.
Built environment – architectural styles, street furniture, etc.

The need to make these kinds of observations demonstrates how the gameplay encourages you to build up a relationship with the environment being shown. When the game begins, you are simply ‘dropped’ into a location with no instructions on where to go or what to look for. Instead, the incentive must come from the player. Every journey is different and personal – which roads to turn down, which details are noticed or ignored, and how those details are interpreted. You perform your relationship with the place into being as you navigate; a kind of cognitive mapping mediated by the screen. By the time you’ve made your guess, having found and deciphered what you’ve seen by travelling steadily at street level, you’ve developed a more deeply affective relationship with the place in question than if you’d just looked at it on a map, or read an encyclopaedia entry.
It is a combination of the instantaneous decisions and reactions the player makes, and their pre-existing paradigm of social norms and individual experiences, that is drawn upon to make sense of the world. By bringing together both of these elements within the overarching timeline of a virtual journey through an unfamiliar environment, GeoGuessr can quite effectively develop meaningful relationships between diverse players and diverse locations, which expand the players’ mental archive of geographical knowledge – albeit within the limits of solely visual communication, and the player’s willingness to keep exploring.
The capacity of the game’s design to grasp and maintain players’ interest is the subject to which I’ll now turn.
Design and technology
GeoGuessr is an example of a game that applies a very simple mechanic very successfully to create a fulfilling play experience. By setting the aim or ‘win condition’ of the game as ‘finding your location’, it inherently encourages players to think about what makes the place in question unique, and pay close attention to detail to find the relevant information.
In the time since its initial release, however, the game’s developers have fashioned new opportunities for players to interact with the photographed environments. In Challenge Mode, the player can limit their guessing time by setting a timer of custom length, forcing them to be more focused in how they filter information. When this mode is selected, the player is also given a URL they can send to friends. This URL gives these others players the same set of five images and same time restrictions, allowing them to compete to get the highest score. From there, it can be fascinating to compare how and why different players construed the information to reach the conclusions they did.
Additionally, using GeoGuessr’s Official Maps, you can narrow down the geographical area from which Streetview images are selected to individual countries, regions, and cities. If you are interested in exploring a particular country or region you have never travelled to, or perhaps want to test your knowledge of a familiar area, these game modes give you the means to do so. By giving the player more agency to choose where they want to explore, the gameplay can provoke a more deeply affecting experience.
In a similar vein, GeoGuessr now gives users to ability to create their own maps and publish them on the website for other players to explore. In the Popular Maps section, players can find numerous user-made, themed maps that allow them to explore the specific types of places that interest them. These include maps with a more generalised remit, such as capital cities, to more niche categories such as Premier League football grounds, or even locations from films and video games.
The downside of this creative feature – that it is behind a paywall. To be able to build your own maps – as well as getting your own personalised pin to drop on the map, and avoiding ads – is a privilege you can only benefit from if you pay $2.99 a month. Of course, many would be happy to pay such an amount for a game they enjoy regularly, and if you’re wanting to create personalised maps for others to try, you’re certainly someone who is more deeply invested in the game. But clearly a paywall can create barriers to engagement, preventing those who cannot afford/do not want to spend that amount of money on the game from becoming more closely involved in the GeoGuessr community.

Despite the possibilities for personalisation, it is important to remember that GeoGuessr is at the mercy of Google with regard to which places are included in the Streetview database, the extent of the imaging that takes place at these locations, and the quality of the final product. Coming across a grainy image – common when looking at parts of the US and Australia – can be a significant source of frustration for players trying to find out more about the place onscreen and make accurate guesses. Moreover, Google updates the images of some locations more frequently than others, meaning that the picture you’re presented with can be several years old. It is not something that can be easily resolved by the game’s developers; though one thing you can do at the end of each round is to give it a rating out of five stars, which then goes on to affect the likelihood of that location being chosen for future players.
One longstanding issue that affects which images are selected in GeoGuessr is the apparent over-representation of some countries compared to others. It is very common to be dropped, at least once in a game, onto a long, straight road in an almost-deserted landscape in Russia or Australia, with little or no locational information to guide you.         
This effect is down to how the game’s algorithms determine which images are shown to players. Countries with a large surface area that Streetview has captured extensively, such as the aforementioned countries, as well as the USA, Canada, Brazil and Scandinavian states, appear more frequently in Standard Mode than their smaller counterparts. While the customised maps are one way around the problem, this matter hints at the power of algorithms in determining which types of information we are exposed to, creating filter bubbles. The developers of GeoGuessr have openly discussed the balance of locations represented many times, and their aim to continue improving this aspect of the experience. But it will always be an intricate balance to ensure a wider selection of places to explore without unduly minimising the presence of other countries, cultures, and types of environment.
Design decisions ultimately have a significant impact within the fields of power that shape representations of places in digital games like GeoGuessr, and consequently how we come to understand them. Of course, it is simultaneously these very same algorithms that make looking through images of such varied environments to be an engaging experience. However, with the over-representation of certain kinds of landscapes, and the reliance on the effective (though repetitive) singular mechanic of finding your location, it is more debatable whether the gameplay can retain player interest continually, for more than a few hours of time spent in-game.
WIth the game’s reliance on resources produced externally – Google Streetview images – GeoGuessr is a particularly thought-provoking case study for looking into the range of power relationships that influence how landscapes are represented in games, and through digital media more generally.


For academics, designers, and others with an interest in how we understand what places are when apprehended through digital technology, it will be instructive to see how these relationships are negotiated as GeoGuessr evolves in the coming months and years. Notably, in November, the game’s developers took to Twitter to ask players for ideas on new features to include, and how the existing gameplay could be improved. There has since been a considerable number of suggestions that, if implemented, could substantially change how the game is played.
In the same way that the evolution of Google Streetview has, exploring environments virtually through GeoGuessr provokes questions about what it means to locate and be located in the digital age. Where the game has been most innovative is in turning this relatively new medium of representing place into an experience that, through its very design, enables players to learn about unfamiliar places by taking their own journeys through series of images, making discoveries and interpretations based on their own perceptions. Where GeoGuessr has limitations, these tend to lie with the Streetview images themselves, and how they are produced, rather than the gameplay. However, the way images are assigned via the game’s algorithm is a pertinent consideration to make when considering how well the game encourages personal discovery and learning about locations.
With forthcoming improvements seemingly on the horizon, it will be intriguing to see how the game’s design branches out from the initial concept, and the effects of future changes on how players make sense of the screened representations of places they encounter through the game.