Yea, solid list Travesty, although a lot of those requests will require access to data that I'm unsure would be given to an intern. Tinder can do lots of interesting data analysis, but they choose not to do it. Sharing data analysis is what made OKCupid popular early on, with their OKTrends blog posts. Tinder's secretive policy probably won't change, especially since they're owned by the same people who presently own OKCupid and put a halt to their OKTrends, IAC/Match. Tinder's owners are not fans of data-mining and sharing.
Having said that, here is the easiest and most practical bit we could learn and employ about the algorithm:
1a. How is the lineup decided?
By this I mean, do guys who swiped 10 hours before get placed on top? Or is it the guys who swiped just 5 minutes before she started?
1b. When are girls most actively swiping
What time of day, per weekday, are girls most active?
1c. Given the answers to 1a and 1b, what is the most beneficial time for a guy to go on a swiping session?
Say girls are more active at 7pm on a Sunday night or whenever. Would it benefit me to have swiped right before her session since the algorithm dictates the more recent swipers go on top? Or would it benefit me to go right after she's done?
1d. How long does a girl's swiping session usually last? 10, 100, 1000 swipes? How imperative is it that I be within a certain range of her list?
The information needed to answer these questions is relatively tame and general info about the algorithm, that doesn't require data mining. If you come away with just this, that would be a victory. Even if the answer is that they're ordered randomly.
The next few requests, though, really require ninja data-mining and statistics skills.
2. What are the gender ratios in different cities, and at different mile radii within a city?
Self explanatory. But I am envisioning a Choropleth map where you can interactively zoom in to a city and check where there are more males and females to like a 5-mile resolution. This would change where a lot of people live, not just what states or cities, but within which areas in a city. Making them is relatively straightforward using software, or linking the data yourself using the D3.js javascript library. If you just get a spreadsheet with two columns, gender and GPS coordinates, you can build something like this interactive map:
I'm not saying this is feasible or realistic, but hey, this is my wish list!
3. What ratio of swipes (for girls) end up in an exchange of numbers? What is the average number of messages exchanged before this happens? What is the highest success rate, 3-4 messages exchanged? 19-20?
These questions sound easier than the above, but actually require acquiring more sensitive user data... what they're actually messaging.
4. Are there any taglines that rise to the top with regards to success rate/number exchange?
As Travesty referred to, does listing your height, a good instagram, etc help?
5. Picture trends
This is actually probably feasible, just going on anecdotal evidence. You can ask around for this, as Travesty's questions are really good. I would add to that the importance of photo quality. And I don't mean physical quality, I mean image quality... that it's obvious it's from a DSLR and is as flattering as possible. Does photo quality up the success rate of average looking dudes?
6. How active are girls when they visit cities outside their own?
Wouldn't be surprising in the least as we all know girls will slut it up when no one's looking. But it would be satisfying to have Tinder confirm that bitches are more swipe happy when vacation dick is on the line.
Mostly this shit just sounds fun man, I would love to work at a place like that, or OKCupid. I'll temper my expectations a bit since they'll probably take the necessary precautions to not let you get access to all that juicy data.
Having said that, here is the easiest and most practical bit we could learn and employ about the algorithm:
1a. How is the lineup decided?
By this I mean, do guys who swiped 10 hours before get placed on top? Or is it the guys who swiped just 5 minutes before she started?
1b. When are girls most actively swiping
What time of day, per weekday, are girls most active?
1c. Given the answers to 1a and 1b, what is the most beneficial time for a guy to go on a swiping session?
Say girls are more active at 7pm on a Sunday night or whenever. Would it benefit me to have swiped right before her session since the algorithm dictates the more recent swipers go on top? Or would it benefit me to go right after she's done?
1d. How long does a girl's swiping session usually last? 10, 100, 1000 swipes? How imperative is it that I be within a certain range of her list?
The information needed to answer these questions is relatively tame and general info about the algorithm, that doesn't require data mining. If you come away with just this, that would be a victory. Even if the answer is that they're ordered randomly.
The next few requests, though, really require ninja data-mining and statistics skills.
2. What are the gender ratios in different cities, and at different mile radii within a city?
Self explanatory. But I am envisioning a Choropleth map where you can interactively zoom in to a city and check where there are more males and females to like a 5-mile resolution. This would change where a lot of people live, not just what states or cities, but within which areas in a city. Making them is relatively straightforward using software, or linking the data yourself using the D3.js javascript library. If you just get a spreadsheet with two columns, gender and GPS coordinates, you can build something like this interactive map:
I'm not saying this is feasible or realistic, but hey, this is my wish list!
3. What ratio of swipes (for girls) end up in an exchange of numbers? What is the average number of messages exchanged before this happens? What is the highest success rate, 3-4 messages exchanged? 19-20?
These questions sound easier than the above, but actually require acquiring more sensitive user data... what they're actually messaging.
4. Are there any taglines that rise to the top with regards to success rate/number exchange?
As Travesty referred to, does listing your height, a good instagram, etc help?
5. Picture trends
This is actually probably feasible, just going on anecdotal evidence. You can ask around for this, as Travesty's questions are really good. I would add to that the importance of photo quality. And I don't mean physical quality, I mean image quality... that it's obvious it's from a DSLR and is as flattering as possible. Does photo quality up the success rate of average looking dudes?
6. How active are girls when they visit cities outside their own?
Wouldn't be surprising in the least as we all know girls will slut it up when no one's looking. But it would be satisfying to have Tinder confirm that bitches are more swipe happy when vacation dick is on the line.
Mostly this shit just sounds fun man, I would love to work at a place like that, or OKCupid. I'll temper my expectations a bit since they'll probably take the necessary precautions to not let you get access to all that juicy data.