Pokémon Go has been a phenomenal success, becoming the biggest US mobile game ever, with over 10% of all Android devices in the country installing it, an attach rate higher than that of Candy Crush’s 8.7%, the former king [3]. A few days after its limited rollout, Nintendo’s shares began to soar, almost doubling its market capitalization in a matter of days [4]. Considering the impact of the game and how popular it has become, I’m pretty confident you have already heard about what all this fuzz is about. But in case you decided to take residence in a cave for the past months, let me recap a bit on what Pokémon Go is.

Pokémon in itself is anything but new. The franchise was created by Satoshi Tajiri in 1995 and is centered on fictional creatures called “Pokémon”, which humans, known as Pokémon Trainers, catch and train to battle each other for sport [5]. The franchise first appeared on Nintendo’s Game Boy console and it has received more than a dozen follow-ups. It is worth noting that, in itself, Pokémon already was a tremendous success. By March of this year, the Pokémon franchise and its spin-offs had already sold about 279 million units worldwide [6].

It is therefore understandable that greedy people thought of bringing the Pokémon fever to mobile devices. Since Nintendo, the parent company and sole owner of Pokémon, has almost no experience developing games for mobile, they partnered with Niantic Labs, an American software developer with more expertise in the area. The result of this partnership was Pokémon Go, a location-based game developed around the idea of turning the users into real Pokémon trainers, who have to capture and train their own creatures. The game features a real world map, where you can see the locations of different Pokémon. In order to capture them, you need to move close to where they are. Once you have approached them, with the use of augmented reality features you can see them in your surroundings and use Poke balls, special ultra high-tech devices that can catch them. Pokémons appear almost randomly around the map and they seem to change their home rather frequently, so you better get them when you see them. This obviously results in hilarious episodes, like the Central Park gathering around a rare Pokémon.

Pokémon Go was officially released on July 6th. Besides the sheer amount of downloads, many other statistics confirm the impact it had in the mobile ecosystem. According to app analytics firm SimilarWeb, Pokémon Go had more daily active users than Twitter, Netflix, Spotify or Pandora [7]. The same firm details that users spend on average 43 minutes a day playing Pokémon Go. This is an outstanding number, even bigger than the average time of use for WhatsApp or Instagram [7]. It has also been a hit in social media, generating over 15.3 million tweets worldwide during its first week of release, almost 4 million more tweets than “Brexit” generated on the week of the referendum [7]. Google searches for the term also spiked drastically in the release week: just try searching Pokémon Go in Google to find over 30 million results. For comparison, looking for Theresa May, the new British prime minister, yields about half of that [8].

Yet more interesting than the statistics about the game are the statistics inside the game. The design and creation of Pokémon Go is based on huge amounts of data and its success is another proof of the reach and potential of Big Data, statistics and analytics technologies. Actually, it is not surprising that Pokémon Go is founded on geo analytics, as using statistics lies within its developers DNA.

Niantic Labs, the developer of Pokémon Go, was formed in 2010, as an internal startup within Google. Yes, that “always using statistics” Google. Its founder is no other than John Hanke, an entrepreneur that in 2000 created a startup called Keyhole. His company was focused on what was then called “spatial visualization”. Just four years after its creation, Google acquired Keyhole and transformed it into what we now know as Google Earth [9]. In the meantime, John Hanke came to be the leader of the Geo Team at Google. For six years, he supervised the development of all the mapping apps Google provides. He and his team shaped and devised Google Earth, Google Maps, Street View, among others. In 2010, Hanke left the leadership of the Geo team to start a pet project, a new startup to experiment with new and risky geo location experiments: Niantic. The name comes from a whaling ship that got dragged on shore in the San Francisco area during the gold rush [9]. The company, powered by all the knowledge, learning and datasets gathered by the head of the Geo apps from Google, started developing location based-apps and games such as Field Trip and Ingress [9]. Due to the use of advanced geo statistics, their location-based games became rather successful, so Niantic came to be a natural choice for Nintendo in order to leverage its popular franchise in mobile devices.

Coming from Google, Niantic has an enviable statistics pedigree. The use of modeling and statistical geo analysis is seen almost everywhere in Pokémon Go. For instance, one of the key components of the game are Pokéstops and Gyms, predefined locations where you can socialize with other players and obtain useful items to aid in your “catch ‘em all” quest. These locations aren’t random at all: they appear only in logical, realistic spots for gathering and socialization. Parks and reunion centres are usually where Gyms are located. With the help of all the crowd sourced data gathered through their previous games, along with an ever growing geo locative dataset, they were able to deploy logical and useful locations for the social aspect of the game [10]. The players know that parks, shopping centres and historic monuments tend to be good spots for Pokémon trainers, as the developer intended them to be.

Even the random location of Pokémons is not completely random, but data-driven. There are several “species” of Pokémon, each with certain “behaviour” that makes them prefer certain areas and surrounding elements than others. It has been reported that aquatic creatures are more likely to appear near lakes, rivers or aquatic reserves. Others tend to be exclusive to parks or green areas. This “clustering” of species within its native environment is also the result of statistical methods that Niantic has developed to understand the worldwide landscape [10].

But not all analytics in Pokémon Go are done by the developers. Being so popular, the game has attracted a few data specialists as well, some of them going as far as developing a large compendium of statistics and tools for the game. Meet Pokelyzer!, a Tableau based Dashboard that allows you to review the locations where Pokémons have appeared to other players [11]. The developer of this Dashboard even added tools to filter Pokémons by rarity or by individual types, as a means to help Pokémon trainers improve their collections. It is quite interesting to identify the kind of data-driven patterns that we discussed where some creatures agglomerate their locations around rivers and other popular spots.

With statistics so deeply involved within the algorithms that made Pokémon Go the hit it has become, it is appropriate to say that the game owes its success to the smart use of analytics. As such, it is nice to know that Pokémon Go is returning the favour to statistics as well. With the current popularity of the game, there is a growing interest in statistical and data mining tools. Considering the impact this app had in its parent company, it is just a matter of time before others start to use data to create even better apps. I know, some will just copy the whole thing (there are some really nice Pokémon Go clones right now [12]), yet some others will find new and more innovative ways to improve games, apps or services with statistics [13] [14]. You won’t need to wait much to hear of a new and sophisticated use of analytics in technology, maybe originated from the growing interest in statistics that Pokémon Go has brought to the table. In the meantime, don’t forget to thank statisticians for the next time you catch a mermaid cat, or whatever Vaporeon happens to be.

References

[1] Vaporeon. Pokémon Official Pokédex Website
http://www.pokemon.com/us/pokedex/vaporeon

[2] Esto es lo que ocurre cuando ‘aparece’ en medio de la noche un Pokémon especial en Central Park. Verne Website – El País, El Periódico Global. (July, 2016)
http://verne.elpais.com/verne/2016/07/16/articulo/1468684654_324197.html

[3] Pokemon GO is now the biggest US mobile game ever. GamesIndustry Website (July, 2016)
http://www.gamesindustry.biz/articles/2016-07-14-pokemon-go-is-now-the-biggest-us-mobile-game-ever

[4] Pokemon GO drives Nintendo market cap past Sony. GamesIndustry Website (July, 2016)
http://www.gamesindustry.biz/articles/2016-07-19-pokemon-go-drives-nintendo-market-cap-past-sony

[5] Pokemon. Wikipedia – The Free Encyclopedia (August, 2016)
https://en.wikipedia.org/wiki/Pokémon

[6] Pokemon franchise crosses 200 million units sold. GamesIndustry Website (July, 2016)
http://www.gamesindustry.biz/articles/2016-03-17-pokemon-franchise-crosses-200-million-units-sold

[7] How Pokemon Go took over the web. BBC News Website (July, 2016)
http://www.bbc.com/news/technology-36824737

[8] Pokémon Go gana al porno y al Candy Crush. Verne Website – El País, El Periódico Global. (July, 2016)
http://verne.elpais.com/verne/2016/07/14/articulo/1468484582_300594.html

[9] Markowitz, Eric. Inside the Mind of Google’s Greatest Idea Man, John Hanke. Inc. Website (Dec, 2012)
http://www.inc.com/eric-markowitz/inside-the-mind-of-googles-greatest-idea-man.html

[10] Augur, Hannah. Pokémon Go and Big Data: You teach me and I’ll teach you. Datachonomy Website (July, 2016)
http://dataconomy.com/pokemon-go-and-big-data/

[11] Finding Hotpots for Locally Rare Pokemon Using Tableau. Whack Data Website (July, 2016)
http://www.whackdata.com/2016/07/27/finding-locally-rare-pokemon/

[12] Millward, Steven. The top game in China right now is a Pokemon Go clone. Techniasia Website (July, 2016)
https://www.techinasia.com/pokemon-go-clone-china

[13] What could the rising interest in Pokémon Go mean for Big Data? Telecoms Website (July, 2016)
http://telecoms.com/opinion/what-could-the-rising-interest-in-pokemon-go-mean-for-big-data/

[14] Reaney, Matt. What Has Pokemon Got To Do With Big Data? KDNuggets Website (July, 2016)
http://www.kdnuggets.com/2016/07/pokemon-big-data.html