Modern, Internet-based social networks contain a wealth of information about each member. An integral part of an individual’s online profile is their Volunteered Geographic Information (VGI) such as a user’s current geographical location. Social network members in different cities, countries, or continents engage in different activities due to accessibility, economy, culture, or other factors. The work here focuses on data mining separate groups of social network profiles according to their geography in order to discover information about a place. This results in keywords associated with a specific location and provides an automated way to describe a place in an up to date fashion based upon its current local residents. Location-Based Social Network (LBSN) profiles from four different places are analyzed here and the results are presented as they relate to space, time, and activities.
TopIntroduction
Digital social networks have seen tremendous growth in contemporary times. Seemingly ubiquitous availability of the Internet has led to billions of social network profiles that each present a virtual representation of a real human being. Some online social networks focus on allowing users to communicate with existing real-world ties while others afford new friendships based upon particular interests. The geographic breadth is intense as individuals from all of the world’s continents partake in membership of one or more online social network(s). Although the popularity of these social network services has risen only in the previous decade, their influence is vast. Online communities allow for easy, efficient communication between many people, which results in emergent events such as flash mobs of pillow fights in San Francisco or political uprisings and the restructuring of governments in the Middle East.
In this chapter the following research questions are addressed:
- •
How can web crawling techniques be used with location-based social networks to create different data repositories according to geographic area?
- •
What role can data mining play in searching for useful information within social network profiles?
- •
How can social network data be visualized in order to demonstrate the importance level of concepts related to space, time, and activity?
The work here aims to address the challenge of data mining location-based social networks in order to discover useful information about different geographic areas. The next section provides a background of relevant literature. The third section presents the approach and methodology used in this chapter. Section 4 contains a discussion of the results and potential future work. The last section completes the chapter with final conclusions.
TopBackground
Social networks have been around since long before the Internet but some of the first Internet-based social networks date back to 1997. One example of these older digital social networks is sixdegrees.com, but many of the older social networks are as active as they once were. Networks have been created for various groups and purposes over the years with many such as Friendster and MySpace having distinct high points of success followed by low levels of membership activity. For years now there have been academic studies of social networks such as to how different roles shape an individual’s activity in the network (Kumar, et al., 2006; Pultar, et al., 2010a). For a general history of social networks with in-depth definitions of the various types the reader is directed to boyd and Ellison (2007).
With years of frequently updated information from all parts of the world, social networks are ripe for data mining and discovery. Data mining (Han & Kamber, 2006; Miller, 2007) is a tool from computer science by which patterns are found in large sets of data that can otherwise be overwhelming. Available through social networks is a wide variety of data about individuals pertaining to topics such as: eating or traveling preferences, philanthropy, identity, and education. See Liu, Maes, and Davenport (2006) for more on how connections between individuals in a social network can be structured by tastes.