Category Archives: Foursquare ~ Keith Walsh

Contextual Awareness Mined from the Data Exhaust

Foursquare’s contextual notifications [1]

As their users were busy playing games and checking-in to cool new places, Foursquare engineers were hard at work. They were mining the data exhaust created by all those check-ins. They were building a recommendation engine that would become contextually aware.

This is where the real value in Foursquare lies. The founders of Foursquare were never running sprints, they were not going to sell out when their user base reached into the millions. Collecting enough data to make relevant predictions of people’s behaviour, would take time. The building of Foursquare was always going to be a marathon. All these games and check-ins, were all the time, adding and updating to a place database which now has over 60 million points of interest. [2] What Foursquare has been creating was a living breathing dataset crowdsourced from its users.

Foursquare has built its contextual notifications from this dataset. Contextual awareness is about providing information based on the context in which you can use it, providing relevant information at the right time. [3] The notifications run in the background where the user doesn’t need to sign-in. It knows when you are in an unfamiliar place and gives you advice and tips only when you need it.

For example, a notification might pop up on your phone with “At Wagamoma’s? People talk about ‘Katsu curry’, ‘Chilli Beef Ramen’, and ‘Get the lunch offer'”. Foursquare knows when you are out of your element and will not give notifications for places near where you work or live. When arriving in a new city a notification could be, “Welcome to Paris! Your friend Geoffrey’s list ‘Hidden Local Gems’ has 6 places nearby”.

To achieve this Foursquare engineers use machine learning techniques, making use not just of location, but time and behaviour of the user and their friends. It uses natural language processing to figure out the sentiment of all the tips and recommendations, understanding which are actionable and useful. [4]

Last month, just hours after Microsoft anounced its new CEO, it invested $15 million into Foursquare. [5] Microsoft are looking for access to Foursquare’s place database to use in Microsoft’s own products and services. In reacting to this news, Foursquare mentioned in its blog, “In the near future, when you use Microsoft devices powered by the Windows and Windows Phone operating systems and products like Bing, places will be enhanced by Foursquare – to provide contextually-aware experiences and the best recommendations of any service in the world.” [6]

It’s not only Microsoft who gets access to Foursquare’s data. The Foursquare API gives access to all of the data used by the foursquare mobile applications. This is part of Foursquare’s long term strategy to become the location layer of the internet. [7] Hundreds of mobile apps now use the Foursquare API, including Uber, Foodspotting, Vine and Instagram. Even after Instagram had been bought out by Facebook it continues to use the Foursquare API for location data over Facebook’s own Place API.  [8]

As Jonathan Barouch, the founder and CEO of location-based startup Roamz has said, “Foursquare’s value is less about the size of its active user base and more related to the reach of its location database. Its API is fast becoming the de facto location layer of the mobile web and touches almost every user of location-based apps.” [8]

But Foursquare now faces some major challenges, how to convince the public that these contextually aware apps bring convenience and not creepiness, and how to persuade their users that they’re giving them something valuable and not simply exploiting their personal data.

~ Keith Walsh

[1] Bilton, R. 2013. “Foursquare brings its recommendation smarts to Android’s push notifications”. VB news. Available at: (Accessed: 13/03/2014)
[2]Hernandez, P. 2014. “Microsoft HereHere Social Platform Aims To Improve Community Engagement”. TechWeekeurope. Available at: (Accessed: 13/03/2014)
[3] Thursday, B. 2014. “Notifications Are Just Getting Started”. Saga. Available at: (accessed: 13/03/2014)
[4] Shaw, B. 2013. “Data Driven Products at Foursquare”. Data Driven NYC 20. Available at: (Accessed: 13/04/2014)
[5] Kelly, C. 2014. “Microsoft Invest $15 Million Into Foursquare, Licenses Location Data”. Forbes. Available at: (Accessed: 13/04/2014)
[6] Foursquare. 2014. “Our crowd-sourced places database has over 60,000,000 entries and 5,000,000,000 check-ins, and one major new partner – Microsoft.” Foursquare Blog. Available at: (Accessed: 13/04/2014)
[7] Olanoff, D. 2013. “Dennis Crowley Says That Foursquare’s API Is Currently Underutilized, Apps That Use Its Location Data Are Smarter”. Techcrunch. Available at: (Accessed: 13/04/2014)
[8] Barouch, J. 2013. “Foursquare’s API Is A Pillar Of The Mobile App Ecosystem”. Techcrunch. Available at: (Accessed: 13/04/2014)


It’s not all fun and games

ImageSome Foursquare badges. [1]

In the early days of Foursquare before a critical mass of users had been established, there needed to be a motivation for a user to use their service. After all, the main reason the first incarnation failed was due to it not gaining this critical mass. The creators realised from their past failure that a new user would not easily take on their product if it wasn’t already being used by their friends. Their solution to this at the time was pioneering and the concept has now become known as gamification.

The term ‘Gamification’ was originally coined in 2002 by a British computer programmer called Nick Pelling. However it wasn’t until around 2010 that the word started to be used for a new concept which sought to implement game mechanics into web and mobile applications. [2] This sought to incorporate the social and reward aspects of games to help solve non-gaming related problems.

The creators of Foursquare were quick to understand the potential of the concept in helping launch their service. This new implementation of game mechanics allowed Foursquare’s early adopters to not depend on another user for intrinsic reinforcement. An idea that was not very intuitive for what was essentially a social media app. Despite this it would be the primary driver of Foursquare’s early success. This is witnessed by Foursquare achieving a phenomenal growth of 3,400% in 2010. [3] Foursquare had proved without any doubt that gamification can affect behaviour and can engage millions of new customers.

Foursquare gamification involves users competing for badges, mayorships and points. To earn a badge a user must check-in to a specific venue or complete certain tasks. The first badge a user gets is called the ‘newbie badge’ and is earned after their first check-in. It is unknown how many badges there is but it is assumed to be in the hundreds. For example, the ‘Hangover badge’ can be earned if the user checks into a venue categorised as a bar after 2am and then checks into a venue categorised as an office or college before 8am the same day. In 2010 an astronaut called Douglas Wheelock earned the ‘NASA explorer badge’ after doing a check-in on the International Space Station. [4]

If a user has managed to check-in to a venue more times than any other user in the last 60 days, then they are awarded mayorship of that venue. This is in condition that the user has uploaded a profile photo. If a business or venue has a foursquare page then the mayor will be displayed on their front page. Many businesses now provide rewards to users who achieve mayorship at their venue.

Foursquare users can earn points each time they do a check-in. Bonus points are awarded for different things such as introducing a new friend to Foursquare, posting a recommendation or checking in to a new place. Foursquare encourages users to compete against their friends for points and to check their position on the points leaderboard.

Recently co-founder Denis Crowley has expressed his disappointment about the perception that Foursquare is just about points and badges. For this reason they are intending to move the game aspects of Foursquare progressively more into the background while emphasising the apps primary functions. [5] In fact the creators of Foursquare have always been upfront about their reasons for using gamification. The function of the games was always to get new users interested, to keep them checking in and to ultimately provide more data for their recommendation engine. [6]

– Keith Walsh

[1] Posttuit. “Coleccionista de medallas”. Available at: (Accessed 06/03/2014)
[2] Gabe Zichermann. 2011. “Gamification by Design: Implementing Game Mechanics in Web and Mobile Apps”. 1
Edition. O’Reilly Media.
[3] Kuo, I. 2013. “Foursquare’s Removal of Gamification: Not a Mistake but a Mature Design Decision”.
Gamification Corp. Available at:
gamification/ (Accessed: 06/03/2014)
[4] Kincaid, J. 2010. “Houston, We’ve Had A Check-In: NASA Astronaut Just Used Foursquare From Space”. Available at: (Accessed:
[5] Crowley, D. 2013. “The Future of Location”. SXSW. Available at: (Accessed: 06/03/2014)
[6] Foursquare. “Foursquare 101”. About Foursquare. Available at: (Accessed: 06/03/2014)

Foursquare: It’s more than just the check-ins.


Figure 1 [1]

“Keep up and meet up with friends on the go” [2] this is the slogan of Foursquare, a social networking app for smart phones. It enables users to share their location and give opinions on venues they’ve visited. It allows the users to know the locations of their friends so if they happen to be both passing near a coffee shop they can meet up for a coffee.

A convergence of new technologies and social attitudes had to take place for such a creation to have been born. At the end of the dot-com bubble social networking sites such as Friendster were beginning to form [3]. The idea was simple; you collect friends like you collect stamps. It was an addictive idea and they soon became extremely popular, however they were limited in their utility. It would take technologies such as smart phones with GPS and projects such as the OpenStreepMap to open up an environment from which Foursquare could be born. Though it was the earlier sites such as Friendster that created the social conditions for which people would become comfortable with sharing the intimate details of their lives. As Mark Zuckerber has said “The rise of social networking online means that people no longer have an expectation of privacy”. [4]

Foursquare was launched in 2009 by Dennis Crowley and Naveen Selvadurai [5]. The idea is similar to an earlier project by Crowley called Dodgeball which was part of his graduate thesis at New York University [6]. Google would later acquire Dodgeball but then decided to discontinue it in 2009. The idea was novel but was severely limited by the technologies available at the time. As there were no GPS enabled phones with anywhere near mass adoption it required users to SMS their location to the service. Internet access on mobile phones at this time was extremely basic with little to no images and slow connections speeds. Because of these limitations the service did not gain the critical mass needed to succeed leading Google to discontinue the service in 2009.

This allowed for Crowley to launch Foursquare with the help of Selvadurai who was an engineering graduate of King’s College London. Selvadurai had worked for Nokia and brought valuable knowledge of the capabilities of the newly appearing smartphones. [7]

Revenue streams for Foursquare come largely through advertising. What makes advertisements on foursquare unique though is that they are location-based. It allows businesses to target users who are nearby their premises. [8]

Recent developments from Foursquare include the Foursquare Time Machine which they have developed with Samsung. This creates an infographic to allow the user to visualise their past checkins, favourite places and movement habits. It also makes predictions on their future movements based on their past data. [9]


Figure 2 [10]

Foursquare has a team of data scientists who apply machine learning techniques to the company’s every expanding datasets [11][12]. When a user checks in to a location such as a restaurant or cafe they leave comments and tips about that place, this information is available to Foursquare to offer recommendations to other users. The data collection doesn’t stop there; the app collects phone’s GPS coordinates, surrounding WiFi network strengths, radio frequency IDs and performs triangulation via distances to nearby mobile phone masts. According to data scientist Blake Shaw the primary goal of the data acquisition is to make their app more useful to the user, “We’ll know that you’ll want coffee in two hours so we can suggest somewhere you can get really good coffee” [13].

The company is beginning to make available their datasets to academics. Data scientists from Foursquare are collaborating with economists to research tipping points and virility behaviour. They are trying to observe if people go to a new venue is there a certain critical point above which behaviour then ripples outwards causing that venue to become successful [14]. Perhaps there could be some really useful research coming out of these collaborations that could be helpful in areas such as traffic management and city planning.

– Keith Walsh

[1] Review Trackers. “How to Claim Your Business on Foursquare”. Available at: (Accessed: 18 February 2014)
[2] The Foursquare Blog. Available at: (Accessed: 18 February 2014)
[3] Chafkin, Max. “How to Kill a Great Idea!”. Available at: (Accessed: 18 February 2014)
[4] Johnson, Bobbie. “Privacy no longer a social norm, says Facebook founder”. Available at: (Accessed: 18 February 2014)
[5] Foursquare. “About Foursquare”. Available at: (Accessed: 18 February 2014)
[6] New York University. “Dodgeball”. Available at: (Accessed 18 February 2014)
[7] Selvadurai, Naveen. Linkedin profile. Available at: (Accessed 18 February 2014)
[8] Foursquare. “Introducing Foursquare Ads”. Available at: (Accessed 18 February 2014)
[9] Weber, Harrison. “Great Scott! Foursquare’s time machine visualizes your past check-ins and predicts where you’ll go next”. Available at: (Accessed 18 February 2014)
[10] Ungerleider, Neal. “Foursquare debuts time machine”. Available at (Accessed 18 February 2014)
[11] Li, Tianhui. Linkedin profile. Available at: (Accessed 18 February 2014)
[12] Shaw, Blake. Linkedin profile. Available at: (Accessed 18 February 2014)
[13] Datoo, Sirak. “Why data science matters to Foursquare”. Available at: (Accessed: 18 February 2014)
[14] Li, Tianhui. “A chat about data science and our fun visualizations”. Available at: (Accessed 18 February 2014)