3

Gamify geni.com

Alexander Bernstein

October 13, 2016 14:22

None

Follow

Over the past few weeks I've been obsessing with finding duplicate profiles in the world tree and merging them, though finding merge candidates as a basic user is a bit challenging. I'm surprised how many duplicates I'm finding. Even though they may be just a tiny fraction of the more than 100 million profiles on Geni, it doesn't seem like many users are doing anything to fix conflicts in their tree - either because they don't  know how, or abandoned their tree, or some other reason.

I think, the quality of the tree could be improved much further, if Geni turned finding duplicate profiles into a game, sort of like Facebook does with places editor <https://www.facebook.com/editor/>. Geni would identify possible merges and present them to active users asking them to decide if two profiles are indeed duplicates. Duplicates that got enough votes (10? 50?) could be merged automatically, or reviewed by curator and then merged.  Users would be awarded points or levels for voting, badges for accuracy, etc. This could be taken further by using crowd consensus to resolve conflicting data resulting from the merges, awarding points for linking sources, etc.   

Even if the awared points are just for bragging rights and can't be turned into real value, it would already provide  enough incentive for some people to come back to the site to earn more points and improve the tree in the process. By allowing points to be used for something, like buying one record, would provide even more incentive for more people to improve the world tree.

 

7 comments

• 1
  
  Anne Brannen October 13, 2016 21:01

  Lol. 

  This would indeed be helpful in some ways, but as a medievalist working in the World Tree, the first thing I think of is how many duplicates there are AND how deeply deeply deeply the medievalists would love it if there weren't more duplicates being created.

  Would the game not encourage people to make MORE duplicates? 

  If so, it would make the medievalists and classicists Sad.

  I have to say, though, it's a sort of enchanting idea.  Except for the creating more duplicates part.

• 0
  
  Alexander Bernstein October 14, 2016 07:37

  Seeing "LOL", I feel that I need to expand a little bit more on the concept. Gamification is not a new thing - it has been studied extensively and applied in a wide variety of areas from marketing, increasing productivity in workplace, and even apparently by some terrorist groups. It meshes very well with crowd-sourcing with the most famous example being Waze: a navigation app that crowd-sources real-time traffic information and mapping data from its users (and was bough by Google for $1 billion). You can learn more about gamification by reading its Wikipedia entry: https://en.wikipedia.org/wiki/Gamification 

  You're right: gaming and cheating usually go together. But the problem with cheating you described wouldn't be more of the duplicates that make you so said. The duplicates would go away immediately as people would merge them to earn points. If the points could be used to say buy a single historical record, then it would be a problem for Geni's bottom-line. But there are ways of dealing with cheats - from reducing the incentive to cheat (e.g. no points on merges of profile that you created yourself, or on recently created profiles) to giving curators power to ban them.

• 0
  
  Anne Brannen October 14, 2016 21:57

  I'm so sorry!  My LOL was NOT about the gaming concept, for which I have great respect.

  It's for the situation in the medieval tree.  We are trying to get people NOT to merge "duplicates" in the medieval tree, because they are so problematic.

  Merging duplicates in the medieval and ancient branches of the tree actually makes much more work for the curators who are working in them!

  I was NOT lolling at you.  No, no.

  But it's difficult in the medieval tree, trying to keep up with the mistakes people are putting in the tree -- the information is usually coming in from unreliable sources, which is what is causing the problems.

   

• 0
  
  Alexander Bernstein October 15, 2016 14:37

   Anne, no need to apologize, no offence was taken. After reading your comment , I though that some people when they read the word "game", wouldn't take the suggestion seriously enough (whether you meant in that way or not). So I decided to add a few comments.

  Your worries about the "unqualified" users doing the merges are very justified. But the game can only help in that respect, in my opinion, by letting users vote on suspected duplicates, but not actually performing the merge. Once enough votes are collected and they are mostly in agreement, a merge could be performed by a curator, for example.

  Take a look at Facebook Places Editor, for example of something like that in action <https://www.facebook.com/editor/>. It shows you places in your town and asks to confirm info about them (phone, address, opening hours, etc.) and whether a places is a duplicate of another. You can confirm exiting info, or enter new one. The suggestions you make do not become live immediately, they need to be reviewed by other users. If enough users says that two places are duplicates of each other, Facebook merges them, I'm guessing, automatically. But it doesn't have to be automatic in case of geni.com

  Facebook also keeps track of your accuracy. If you said that a place is a restaurant, but most other users said it's a grocery store, your accuracy goes down. I don't know if facebook actually does that, but one could make votes of users with bad accuracy less significant.

  Options are endless, and may need a lot of tuning. The point is, the power of collaborative tree is in its users being active in improving it. However, it seems most people enter their info and leave. Geni should do more to keep users engaged with the site to actively improve the quality of its world tree.

   

• 0
  
  Jeff, Geni Curator October 15, 2016 15:25

  I've always been a big fan of gamification and I think it could be used in various areas on Geni as a rewarding method for positive activity, be it merging, helping in a discussion, fixing errors, or whatever.  Likewise, negative "points" could be applied to problematic activity.  Sometimes they're tied together, you spend your positive points when applying negative points to another user.  At a certain point, curators could review users with big negatives that might be impacting the world tree and suspend as needed.

• 0
  
  Eric Randol Schoenberg, Geni Curator October 25, 2016 23:27

  Encouraging positive/productive behavior is something I have long advocated for.  I am in favor of making much of the Top 10 List fields tally automatically https://www.geni.com/projects/Geni-Top-10-Lists/11937, with charts by month, year, lifetime.  This would also allow curators to identify users who are very active.

• 0
  
  Yao Liu February 13, 2017 20:32

  That's a nice idea! Geni likens itself to a giant jigsaw puzzle, and we certainly could do better at engaging users to work on the Tree, especially parts that are far off from our immediate family.

  Instead of presenting users with random matches, it may be better to cue them with profiles that are in projects that they follow. Maybe not everyone is aware of it, you could go to the list of profiles of any project, and filter by "with tree matches" to find possible duplicates. It would be great if project collaborators (and maybe even followers) be notified when new matches turn up.

  Taking it further, what if we also get notified of matches that are a few steps from those in projects we follow? (And, to get us engaged, tell us this is say Einstein grandma. This would be useful across Geni.) Would that be too complicated computationally?

  Edited by Yao Liu February 18, 2017 05:30