As of April 2015, one in every eighteen United States citizens are using big data to find a companionship [9]. In the age of online dating, big data analytics has become a major contributor to leading to potential relationship success, because online dating services have to deal with a huge amount of data. As an example, Match.com has collected over seventy terabytes of data on their users [9]. Match.com claims that, with the help of big data analytics, they have created of 500,000 relationships resulting in 92,000 marriages and one million babies being born [9]. This demonstrates that technology and big data are changing the dating game.
5.3.1 Generating and Collecting Big Data
Online dating sites use many methods to generate and collect data about their customers. Typically, most information is gathered through questionnaires [9]. The questionnaires ask for likes, dislikes, interests, hobbies, and so on. The number of questions asked depends on the service that the user has selected. It appears that the more successful sites ask hundreds of questions to get better results [9]. Diagram shown in Figure 6 provided by an article [9] illustrates a simple depiction on how matches are made based on the information provided.
Figure 1: Diagram showing how data is used to make matched
In addition to questionnaires, some sites collect data about customers from social media accounts and online shopping history by asking for user permission to have access to those accounts [9]. This information allows online dating sites to observe the actions of its customers, not only what is filled out in a questionnaire [9]. After the site collects a large amount of data, the information is analyzed; all the data is compiled in a database system including RDBMS and NoSQL databases, and then sifted through using a variety of different algorithms to predict the best matches [9].
5.3.2 Analyzing Big Data
The main objective in online dating is to find accurate matches. However, it is debatable whether big data actually improves the chances of a potential soulmate. Those against big data in online dating claim that there is a high probability that both females and males may unintentionally or intentionally misrepresents themselves [9]. This is a major weakness for online dating sites to overcome. Human error is why many sites are attempting to understand their customer’s behaviors [9]. This is done by obtaining their search history, shopping history, and profiles on social media sites.
Other professionals believe that big data is essential to finding the right relationship. The thought is that big data creates facts, and facts do not lie [9]. As mentioned in the previous section, many sites are using customers’ online behavior to suggest potential matches. These behaviors include where the customer likes to shop, what shows they watch on Netflix, what social media site they perform, and so on. Dr. Zhao from the University of Iowa has created a collaborative filtering system that looks at browsing behavior, in addition to responses from potential matches [3]. Examples of the browsing behavior are where does this person shop online and what music do they listen to. This particular algorithm for online dating works similarly to how Netflix and Amazon recommend certain products [3].
5.3.3 Examples of Big Data in Online Dating
Almost every dating site has created their own algorithms using big data in order to create meticulous matches. Match.com has over seventy terabytes or data while eHarmony has over one hundred and twenty terabytes [9]. The next two paragraphs will analyze big data techniques that eHarmony and Match.com uses to determine a match.
eHarmony:
The datasets created by eHarmony’s algorithm are four terabytes each day [9]. Every piece of information collected by eHarmony is used to determine each likely match for their users [9]. eHarmony currently has different algorithms working together to sort and analyze large amounts of data [9]. In addition to big data, eHarmony also utilizes machine learning to establish over one billion matches daily [9]. The matchmaking system for eHarmony is built in MongoDB which allows matches to be made in under twelve hours [9].
Match.com:
Match.com provides questionnaires that range from fifteen to one hundred questions [9]. Next, points are given to the user based on a variety of predetermined qualifications. For example, how important is it that your potential partner answers this question in a similar way [9]? Once the points have been assigned, users with similar points are matched together. Instead of using big data to create matches, Match.com uses their big data algorithm to discover any inconsistencies within the match. Match.com looks at the customer’s answers in comparison to their actions on the website [9]. If distinct differences are found, the algorithm adjusts the match to create more accurate depiction of the user [9]. In addition, Match.com uses facial recognition algorithm that looks at the user’s previous chosen match to determine physical features the user has favored previously [9].
Tinder:
Tinder is a casual dating site that allows user to make split second decisions to determine if they like a potential match [12]. This mobile application show a vague profile illustrate in figure 7 . The user then swipes right on the profile to match the potential suitor. If the potential suitor also swipes right, a match is made and both parties are alerted [12].
Figure 2: A sample profile from the dating app Tinder.
Recently, Tinder had overzealous right swipe clients. If every user of the application swiped right, it would lower the value of the right swipe overall [12]. To elaborate, users would not take any matches seriously, because every profile will ultimately match one another. To fix this issue, Tinder set a limit of right swipe that users are allowed to have each day [12]. To determine if this change affected their membership, Tinder collected big data on their users that only swipe right. Tinder found that the users conformed to the new rules and did not discontinue their membership [12].
Tinder is currently using a software called Interana to collect data from their clients [12]. Interana is a self service tool that analyzes data by allowing users to input queries [12]. These queries are entered into the database without using complex coding and receive feedback in seconds[ 12]. This is a huge step in big data analysis that typically needs custom SQL queries.