Thoughts as a Foursquare Intern…
As I approach the final weeks of my internship at Foursquare, I feel as though I have learned a great deal about the tech world generally and about Foursquare’s strategic direction. Even after week one, I was using words that had not previously been in my vocabulary to describe my projects and Foursquare’s newest goals to my friends and family. I was fortunate to land a spot on the Creative + Strategy team, which is dedicated to Foursquare’s B2B strategy as well as the creative and business strategy for all of Foursquare’s brand clients. I am working with a terrific group of intelligent and motivated people. Everyone in the office is extremely warm and very clearly dedicated to the success of Foursquare. At the company-wide meeting my first week, the CEO, Dennis Crowley, personally introduced me to the group and asked me to share a fun fact with everyone. (In case you were wondering, my fun fact is that I’m a certified scuba diver!)
Foursquare is not my first internship opportunity, but this internship is unlike any of my previous experiences. Maybe it’s because I feel relaxed in the open and chill environment of the Foursquare office. Or maybe it’s because my coworkers’ trust in me and willingness to allocate important projects to me reaffirms that I am having a real impact on the business. During my first week, my team leader spent an hour teaching me, one-on-one, about the data visualization tool that he recently created in partnership with Code & Theory. A good amount of my time has been spent designing data visualizations through this tool, and throughout this article I am going to give you some insight into how it works.
A little bit of background: Foursquare launched in 2009 as a consumer app that allowed users to check in to places. Five years later, the company made bold steps to unbundle Foursquare’s most popular use cases and created two apps: Foursquare and Swarm. Swarm incorporates the check in and location sharing mentioned above while Foursquare provides personalized recommendations based upon people’s tastes and location. With six years of operations, Foursquare has aggregated an immense amount of location data and a unique understanding of how people move through the world. More than 85,000 developers (including some of the world’s most popular apps) use Foursquare’s first-party place data to power their location services. As the only company with this geo-location knowledge, Foursquare has become the go-to provider for major consumer brands that want to enhance their user experience by baking in location capabilities. When you open Twitter on your phone and set a location to your tweet, the list of nearby places that populates is powered by Foursquare. Microsoft’s Bing and Cortana also source information from Foursquare when people search for places on their platform.
Foursquare believes that what you do in the physical world is the best representation of who you are. While you may research various interests on the Internet, the places you choose to spend your free time ultimately reveal the most about what you value. Using this information, Foursquare works with marketers and advertising agencies to help brands connect with their target audiences more effectively.
Foursquare has built software to better understand how places look on a map by using GPS signals to create ‘place shapes.’ Check-ins at a place aggregate phone signals over time and by leveraging Gaussian probability models, Foursquare’s data science team has been able to create areas of probability within which we can assume that a GPS signal is associated with a certain venue. The areas of probability are comprised of various shapes, rather than a simple radius or square around a point, and represent the ‘place shapes.’ When a mobile GPS signal appears within a venue’s place shape, the individual using the phone is “snapped to place,” or assumed to be at that venue due to high confidence levels. A “snap-to-place” helps Foursquare understand where you are, show you relevant tips and learn more about the types of places you go. In order to visually demonstrate the sophistication of their location knowledge, Foursquare developed a data visualization tool that shows how your phone’s sensors perceive places. This software is brand new and I am very lucky to be one of the first people in the whole world to learn how to program on it. Essentially, the tool pulls location data from the Foursquare database using a unique 16-character alphanumeric code. This code is known internally as the Venue ID, which is associated with a large amount of place-specific data on Foursquare’s servers. The venue dataset contains lat-long information, hours of operation, photos, and much more. It also contains the ‘place shape’ of the location as mapped by phone sensor data. The data visualization tool has created a way to show what these models look like by representing them in code overlaid on Mapbox satellite imagery.
When looking at an actual data visualization model, the white lines, or ‘markers,’ represent the specific check-ins at that location. The amount of markers displayed change in correlation with the traffic patterns throughout the week. In addition, a ‘point cloud’ shows the density of check-ins at various levels of GPS confidence associated with a certain place. The color and height of the point cloud represents Foursquare’s level of confidence that the user was at the specified location. Blue cloud points indicate high confidence check-ins at the venue while pink cloud points indicate lower confidence check-ins. The tool then interprets the aggregated place shape data to create ‘contours’ that estimate the areas of probability that a user is at a place. As the ‘contours’ increase in size, the degree of confidence that a user within that location is actually there decreases.
Below is an example of a data visualization that I created. The data visualization illustrates the models for the High Line, Santina restaurant and Brass Monkey bar. I chose these venues because Santina and Brass Monkey are both located below the High Line, which demonstrates the tools ability to distinguish between venues at the same lat-long location. Hopefully this video will provide with you a better sense of the insights we are able to glean from Foursquare’s first-party data and allow you to better understand the data visualization tool.
This data visualization tool is just one way that Foursquare educates businesses on location technology and provides insight into consumer behavior across various spectrums. Businesses use this information to design a highly targeted advertising approach or to shape general business strategy such as planning retail, field sales or predicting market trends. Learning how to program a tool that is so impactful to Foursquare and other businesses has been an enormously exciting experience. With all of the advancements going on at Foursquare, even since I have begun my internship, I am eager to see where the company goes and how the data visualization tool impacts business.