What algorithms do dating apps use to find your next match? How is your personal data impacting your decision to go on a date? How is AI affecting your dating life?  Find out below.

Technology has changed the way we communicate, the way we move, and the way we consume content. It’s also changing the way we meet people. Looking for a partner online is a more common occurrence than searching for one in person. According to a study by Online Dating Magazine, there are almost 8,000 dating sites out there, so the opportunity and potential to find love is limitless. Besides presenting potential partners and the opportunity for love, these sites have another thing in common — data. Have you ever thought about how dating apps use the data you give them?   

How are dating apps using your data?

All dating applications ask the user for multiple levels of preferences in a partner, personality traits, and preferred hobbies, which raises the question: How do dating sites use this data? On the surface, it seems that they simply use this data to assist users in finding the best possible potential partner. Dating application users are frequently asked for their own location, height, profession, religion, hobbies, and interests. How do dating sites actually use this information as a call to action to find you a match? 

• Natural Language Processing (NLP) looks at social media feeds to make conclusions about users and assess potential compatibility with others. AI programs use this input to look for other users with similar input to present to the user. Furthermore, these programs learn user preferences based on profiles that they agree to or reject. Simply put, the application learns the types of people you are liking and will subsequently put more people like that in front of you to choose from. 
• Deep Learning (DL) sorts through facial features of profiles that you have “liked” or “disliked.” Depending on how homogenous your “likes” are, the variety of options presented to you will change. 

What algorithms are these dating apps using?

Hinge calls itself “the dating app that was designed to be deleted.” It uses a Nobel Prize winning algorithm to put its users together. Known as the Gale-Shipley algorithm, this method looks at users’ preferences, acceptances, and rejections to pair people together. Hinge presents this information to the user with a notification at the top of the screen that lets the person know of high potential compatibility with the given profile. Research shows that since launching this “Most Compatible” feature, Hinge been able to guide its users toward people more suited for them. Research shows that people were eight times more likely to swipe right and agree to a “most compatible” recommendation than the alternative without one. This is ultimately resulting in not only more relationships, but relationships of better quality as well. 

OkCupid’s algorithm uses a similar compatibility feature to match its users together. When filling out a profile for this dating app, users can respond to an extensive questionnaire about their personal traits as well as the traits they are looking for in a partner. For example, someone could report that they are very messy and looking for someone moderately messy. OkCupid would then present the user with potential partners who are moderately messy looking for people who are very messy. The algorithm goes one step further than simple response based matching, it ranks the importance of each trait to pair users as well. This approach must be working because OkCupid was the most mentioned dating app in the New York Times wedding section. 

Not all dating apps use this compatibility approach. Tinder, for instance, relies almost completely  on location and images to suggest potential partners to its users. The other aspect to Tinder’s algorithm is based on a desirability factor. In this case, the more “likes” you get will result in people being presented to you who also get a lot of “likes.” It also works in the opposite circumstance where users who don’t receive a lot of “likes” will be presented with people who also don’t receive a lot of “likes.” As a result, 1.6 billion swipes occur daily on Tinder.

A final example of algorithms in dating apps is how Bumble users can now filter preferences beyond personality traits, professions, and appearances. They are able to filter potential partners  by zodiac signs. In many cultures across the globe, astrological signs have been and continue to be used to measure the compatibility of a couple. Bumble’s AI program takes into account user preferences as well as sign compatibility when presenting a potential partner to its user. Matching zodiac signs is another instance of dating app technology working with user data to create the most compatible matches. The extensiveness of Bumble’s algorithm results in over 60% of matches leading to a conversation. See the chart below for the most popular zodiac signs according to a study of 40 million users by Jaumo. 

Conclusion

AI in dating sites goes beyond the individual’s knowledge of their own personality. It gets to know the users better than they know themselves. By monitoring both user input and user behavior, AI in dating applications truly gets to know the most holistic version of the user. It goes beyond the user’s own notion of themself to reveal truths about the type of partner they are  really looking for. The AI in dating apps aims to reconcile a user’s idealized version of a potential partner with the reality of the types of profiles they like. The trajectory of this revolutionizes the way data will continue to be used in AI mechanisms to help humanity achieve results on multiple platforms, even in dating.

What’s the most romantic city in America? What kind of gifts are people buying their Valentine? Given the size of the online dating industry, it’s no surprise that they’ve started leveraging big data to create better matching systems. The real million dollar question is: can you leverage all those numbers and data points to help yourself?

If you’re single, you may have already considered online dating. Nearly 11% of American adults have tried it, and chances are you may end up on Match, OKCupid, or any one of the countless dating sites. Some try to rig the dating-game from the very beginning. They consider which cities will have the best matches for them, or what line of work will get them the most attention. Plenty of Fish did their own data study and found Portland to be the “most romantic city in the US.” While Michigan was found to have the most romantic singles, Louisiana came in last. They did this using data from 5 million singles. “The most romantic places were determined by the percentage of singles within that region who list interests like ‘romance’, ‘long walks on the beach’, ‘cuddling by the fire’, (and thousands of other romantic phrases) on their PlentyOfFish profiles.”

Now, the problem with data in online dating has already presented itself. Determining something as abstract as “romance” (or even “love”) with data is not easy. Some of the other top romantic interests listed on profile included:

• Holding Hands
• Bubble Baths
• Romantic massages

Given the fact many of us would never list these things on a profile begs the question whether PlentyOfFish found the most romantic states, or simply the cheesiest. Many are skeptical about data in the dating industry, and stress that it’s data quality that matters most. Christian Rudder, one founder of OKCupid, explains that one of the biggest hurdles faced in the industry isn’t just finding the algorithm, but finding the right data.

“My intuition is that most of what users enter is true, but people do misunderstand themselves,” he says. Even if a couple seems to match mathematically, there’s a level of superficiality and less-than-perfect information companies need to peel back in order to get a realistic match. One study from Berkeley found that “81 percent of online daters reported inaccurate information about their weight, height, or age,” and that’s likely not even on accident! For example, while someone might list “classical music” as an interest, they really only mean they like it in a vague, theoretical way. Analyzing their personal playlists might prove that they, in fact, care little about the genre. This brings online dater’s to the two important rules:

First, be honest in questionnaires. They may be frustrating. You may want to sound more interesting. You may be thinking, “just show me the matches, already!” These algorithms, however, can only work with the data that’s given to them. Giving flawed information will mean more flawed dates. Second, when possible, connect to other outlets. If users are willing to give permission for companies to scan their Spotify, Netflix, Facebook or search histories, a wealth of far more reliable data can be used. Good algorithms won’t just match you based on mutual friends, or whether you both like Downtown Abbey. They can create several new models for finding matches.

One unexpected method is by comparing users against their competition. If two users seem to have similar music taste and keep chatting with similar people, data from one individual may help generate matches or information for the other. This can also help engines determine just how desirable your own profile is to other users. Algorithms can also decide how attractive your profile is by comparing it to similar users and their popularity—which does sound a little scary.

Dating Data Must Be Used More Creatively

Rudder of OKCupid revealed that there is surprising information data analysis has returned. By compiling data of OKCupid members who ended up in relationship via the online platform, they found three questions most first dates agreed on:

1. “Do you like horror movies?”
2. “Have you ever travelled around another country alone?”
3. “Wouldn’t it be fun to chuck it all and go live on a sailboat?”

The seemingly innocuous questions reveal much more about personality and life trajectory than dozens of useless data points. Amy Webb’s TEDTalk, titled “How I Hacked Online Dating,” has almost 4.5 million views. This is because people not only find the topic interesting, but they likely have had similar, perhaps negative, experiences in online dating. Users are asked questions that, while useful, can’t encapsulate a person as a whole. What she did was reverse-engineer the system and create her own data points to find Mr. Right. She used 72 data points to find a match…and it worked! She had great success finding a sea of quality fish. Unfortunately, they didn’t like her back, due to the way she, herself, had presented and assembled her profile. That led her to study what made other users likable and popular. The results ranged average message word counts (97), to average time between communications (23 hours) and, of course, the photos. The problem wasn’t a lack of data—it was just the wrong data.

How should a data-nerd show their Valentine they care?

By leveraging data. The National Retail Foundation is more than happy to analyze where shoppers will be spending their money on Valentine’s Day. In the past, they’ve found discount and department stores to be major destinations. They even pinpointed the average spending of those celebrating Valentine’s would be $87.94 exactly in 2015. They know folks aged 45-54 were the biggest spenders, and that men were creating more music playlists for the occasion. This is the kind of data that informs, but doesn’t necessarily tell lovers what to do.

Just like Michigan shouldn’t be deemed the most romantic state because of seemingly “romantic” statements, data can help inform lovers on how to get a little more creative. Statistics from StatisticBrain show some 198,000,000 roses will be purchased for Valentine’s Day. The most given gift is cards, followed by candy, dinner, flowers and jewelry—gift cards even made the list. Flower purchases at Proflowers.com and 1800flowers.com peaked enormously last year on February 13th. The same occurred on 123greetings.com, an e-greeting site. Lovers of the world, data can be leveraged to find love and to keep. The key is to use it creatively and correctly.

image source: Online Dating University

Like this article? Subscribe to our weekly newsletter to never miss out!

Follow @DataconomyMedia

We recently had the opportunity to sit down with Kris Hammond, the Chief Scientist for Narrative Science. Narrative Science focuses around automating text generated from data, turning raw data into insightful accounts. Hammond has spent over 20 years working in and developing the AI labs at the University of Chicago and Northwestern University, making him uniquely placed to offer perspectives on the past, present and future of AI. In the first part of our discussion, we discussed the technologies which will shape the future of machine learning; in this installment, Hammond discusses the future of AI, and whether or not robots could actually wipe out humanity and steal our jobs.

When we talk about AI, almost anyone you talk with will say that they think that AI image- the genuine artificial intelligence that is building a system as intelligent, if not more intelligent than a human being- is simply not feasible or possible. Unless we start talking about machines killing us, and then the response is “Oh my god, we have to be terrified of this”.

I think the reality is that we have complete flexibility in terms of building the things that we’re going to build. In order to be a true AI, the future of AI is going to have a goal structure associated with it. Really, all you need to do is make sure that one of the higher priority goals is don’t kill everybody. I know Elon Musk is a very present figure, very smart man, but what I’m worried about existential threats, I’m actually a little more worried about New York being underwater in 30 years. That worries me alot more than the vague possibility of AI which decides to hunt us down and kill us.  In fact from a Narrative Science point of view, we look at what we do when we think, what’s Quill going to do? Explain someone to death? Because  that’s what it does: explaining things.

So I think when we get a little further down the line, and we get closer and closer to what looks like a genuine, complete AI systems, that’s when it’s time to consider, “Okay, what are the constraints going to be?” But the notion that we should start regulating now, as Musk suggests? I think that’s absurdist. There is no point in regulating something that is a glint at this point in people’s eyes. Now, I actually do believe that we will have complete AI. I believe that people are causal beings and that AI and computers live in the same causal environment, and we will have machines that are as- if not more- intelligent than we are. Maybe in my lifetime.

But it’s not time to worry about killing sprees quite yet. Although my concern is that right now a third of the marriages in United States at least, were the result of online dating. Which means that there are algorithms out there that are actually determining the breeding habits of people in the United States. If I were an AI, I wouldn’t blow everyone up. I’d just insert myself into that process and just make sure that system matched up people who were nice and calm, and make the entire species calm for the  rest of time.

For a lot of people historically, AI has meant ‘killer robots’. I understand that. But nowadays, there seems to be this huge focus on AI stepping in and taking over jobs, and automation in general. And most for most of us, there’s still a focus on the blue collar side, but I think that there’s a growing awareness of the white collar side.

I think the reality is that AI is not going to take over jobs; it’s going to take over work. If you look at the work that Watson’s taking on, that Narrative Science is taking on, it’s the work that’s not particularly interesting or enjoyable for people. Having Narrative Science step in to look at the data and do the reporting means that the people who were doing that reporting can step away from doing commodity work and they can actually start working on what a data scientist or an analyst should be doing. They can focus on more speculative work, more discovery work, exploratory work against that data, to find new things instead of reporting on the things they have already found.

I think for AI in general, the goal is not to make the machine smarter and destroy us, but to make machines smarter and as a result, put us in a position where we no longer have to deal with the machine, as an unintelligent device which requires frequent input and supervision. We can deal with the machine as a partner, whose job is to make us smarter. We get smarter because it gets smarter. Because who in the world wants to actually look at a spreadsheet, or figure out what’s going on in the visualization, or go to massive textual data to get the answer to a question? No one wants to do that. As the machine takes more and more of that on, our lives become more human.

And so, AI moving forward is part of the process of actually more deeply humanizing us in our work, in our lives, in our thinking. I think there will be a moment where we embrace that finally, but I wish we could get to it. Understand the excitement of having intelligent partners, whose job is to help us and help move us forward, and to give us more of what it means to be human.

Follow @DataconomyMedia

(Image credit: Saad Faruque)

I was recently dared to register for an online dating account, and just to prove that I’m not actually an asexual library hermit whose love affairs will only ever be with Tolstoy and Tennyson, I did it. Unfortunately, my love affair with OkCupid was short-lived; I was about to exit the site altogether before the button “Boost Profile!” caught my eye. Intrigued, I clicked on the category for ‘10 Boosts’, and this showed up:

“With extra momentum in our algorithms, we’ll show you to more people, faster. Pay an extra £1.05 per boost!”

Wow. I can now pay to have an algorithm tweaked in favour of my chances at love. Talk about technology taking leaps and strides. But all jokey cynicism aside, my mini-foray into online dating got me thinking about how big data facilitates for easier access to sex, love and matchmaking, and whether this heralds a positive turn for human connectedness. Does it foster a hook-up culture, or does it actually help create more modern-day fairytales? Admittedly this is a subject done to death by data scientists, hip journos and most recently, the matchmaker extraordinaire himself – Christian Rudder, cofounder of OKCupid and author of Dataclysm: Who We Are (When We Think No One Is Looking).

But I’m here to specifically address two common concerns about the role of big data in online dating, and hopefully debunk them in my arguments below.

First, we must establish why big data works better than personalised matchmaking services. Remembering what Viktor Mayer-Schonberger said, big data is something that “transforms figures into something more probabilistic than precise. We often need to embrace messiness when we increase scale”.

In the context of online dating, ‘messiness’ is a virtue, because it implies that one can meet more, rather than a limited (albeit meticulously selected), pool of potential partners. In light of Banko and Brill’s discussion on the trade-off between corpus development and algorithm development, Plenty of Fish’s business model is successful mainly because Markus Frind, the company’s CEO, focuses on its corpus development by ever-expanding its database (i.e. increasing the number of sign-ups). Yet that’s not to say algorithm development isn’t just as important. In fact, while the former may be the best way to boost the total membership count (As of 2013, Plenty of Fish has 55m+ members, with OKCupid and Chemistry coming in a distant second at 30m), it is the latter that will convert free members into paid subscribers and keep users coming back to the site for more, hence sustaining a steady revenue flow. But is online dating really such a win-win situation for both businesses and users?

Currently, there are two main arguments against relying on big data as the new matchmaker of our age, and here’s why they are problematic:

1) Big data ‘commodifies’ romance and turns dating into a business transaction

In Dan Slater’s Love in the Time of Algorithms, the author suggests that the idea of romance as “a supernatural force existing outside of conventional social institutions” only gained currency with the rise of Romanticism in the late 18th century, which marked a tectonic shift in the way people perceived courtship and marriage as individual, rather than social, acts. The human search for a companion became a hero’s quest for the sublime, and with this elevation in concept came idealistic expectations about finding the ‘one, true, everlasting soul mate’.

Enter online dating, and the digital generation is now being asked to cast this mindset aside, and to instead see mating through a more neutral, pluralising set of lens just as they would of shopping or decision-making in general. The anti camp would cry foul at how big data reduces everyone to the sum of their profile, and that individuals are now no better than items stacked on supermarket shelves, catalogued and slotted into neat thumbnails ready for the prospective dater to ‘pick ‘n mix’. This goes against the principle of ‘fate’, they say, and the very idea that love should be engineered is sacrilege.

But what if I were to argue that different times call for different measures, and for millions of urbanites who work round the clock and have little opportunity or time to venture beyond their typical environment, targeted search is likely to be a lot more effective in spotting ‘the one’ among many. Let’s take a more radical step: What if I were to de-romanticise romance and say that finding a romantic interest is really not all that different from finding an online tutor or a hair stylist? In all cases, we’re concerned with finding a suitable person who we’d then peg our hopes on and (hopefully) develop an interpersonal rapport with. Obviously the depth and duration of such connection will differ, but the goal remains unchanged: we want to find that eligible someone who will ‘fit’ our bill, and the use of big data allows us to do so “within any given pool…matching up potential pairs” in a way that is “as efficient as possible, getting you what you say you want, or what the data reflects that you seem to actually want”. If anything, it’s catalysing and increasing human connection on an unprecedented level in history. The Romantics will hold on to the principle of data and technology having no place in the ‘soul-searching’ process, but in this day and age, this sort of Luddite argument can no longer stand its ground against the power of efficiency and effectiveness. 

2) With more choices, people will treasure their relationships less

According to Slater, most of the dating executives interviewed for his book agreed that “the rise of online dating will lead to an overall decrease in commitment”. This is not only a crude appropriation of Barry Schwartz’s ‘Paradox of Choice’ theory, it also tars all of the users with a cynical and generalising brush.

To illustrate with an analogy: say I were presented with 10 doughnuts of various flavours, and because I only really like chocolate, I’d naturally pick the chocolate-flavoured one. Would I then be any less committed to eating what I picked because I had to forgo the other strawberry- lemon- or vanilla-flavoured doughnuts, all of which I was never interested in from the outset? The same logic applies even more so for dating, because once I’ve established that someone fits my bill, virtually no one can replicate this person. Big data may open up more choices for us, but this increase isn’t a de facto license to philander, and if we think about the increase in options not as an end in itself but rather as a more effective means of narrowing down that ‘special’ someone from a larger pool of possibilities, then this claim holds very little weight.

Moreover, easier access to more profiles doesn’t necessarily mean easier access to the owners of these profiles. What if they don’t respond to your messages/just aren’t interested/think that you’re weird for listening to Frank Sinatra? There are still multiple hurdles to overcome before one can successfully communicate with another user, let alone develop an actual relationship with him or her. As Alex Mehr, cofounder of Zoosk, puts it aptly, commitment is ultimately “a personality thing”, and “online dating does nothing more than remove a barrier to meeting”. The bottom line is, unless you’re an unfeeling douche or a womaniser at heart, it’s really not that easy to be cavalier about relationships just because the possibility of more romantic options is suddenly made open.

Ultimately, choice can be a blessing or a curse, and what will come of online dating depends on how we use it. So instead of seeing data as an underground Pandarus manipulating the course of love, perhaps it’s high time we openly embraced it as the digital Cupid of our era.

Follow @DataconomyMedia

Jennifer Chan is currently an English Literature finalist at the University of Oxford. She has taken up many editorial roles, having been the online editor of The Oxford Student, one of the two major newspapers on campus; the economics editor of The Oxonian Globalist, an academically-oriented journal on international affairs; and the editorial assistant of the 2014-5 Oxford University Careers Guide. Jennifer is passionate about exploring the relevance of Big Data to the ‘every man’, gauging the function that data analytics may play in the equalisation of education, and familiarising the public with this concept through examples drawn from daily life.

(Image source: Flickr)