selbyk/linkedin-economic-graph-challenge

LinkedIn Economic Graph Challenge

Proposal submission to The LinkedIn Economic Graph Challenge.

Here are the rules, and my idea is below.

My idea stems from Sandy Pentland’s lecture on Social Physics: How Good Ideas Spread, one of the many great Talks at Google. Watching Social Physics: How Good Ideas Spread is required to understand the mechanisms I am trying to incorporate. I do not have the time or resources to do this alone, but if the right team comes together, we might be able to build a successful system.

A System for Sustained Network Growth While Increasing Network Activity

Building on Social Reward Systems and Their Observed Effect on Network Activity

---

Abstract Requirements

1. Main idea and how it's related to the Economic Graph initiative
2. Expected impact
3. High level approach

I want to help build LinkedIn's vision of economic opportunity for every member of the global workforce by applying social reward techniques shown to double observed network activity in a way that may stimulate the reach of LinkedIn's network while also increasing the activity of users within LinkedIn communities over a sustained period of time, shown to last even after the programs have ended. I think improving LinkedIn’s user invite system by integrating a social rewards system designed to keep invited users engaged in LinkedIn communities of their choosing could help acheive LinkedIn's vision. By distinguishing old relationships, the foudation and growth of communities within LinkedIn can be encouraged by allowing users who meet certain relationship requirements (closer relationships prefered) to opt-in to a social rewards system that works based on the activity within LinkedIn communities of a close peer they've invited to LinkedIn. Limiting the system to only users with previous close connections that can be determined from publicly avliable data, allows the program to feel friendly and engaging rather than annoying, and our methods will need to be tweaked along the way.

Proposal Requirements:

1. Describe your proposal in full details and explain why it's novel
2. Describe the potential impact if the research is conducted successfully
3. Describe the methods and tools you plan to use for your research
4. Describe the type of data you plan to use for your research
5. Describe the institutional support/approvals that you have already received from this research (if any)
6. The main proposal should not exceed 1000 words. You may provide link to supplementary materials below

I believe the users participating in the social rewards system will generate a more sustained increase in network activity than users invited who do not participate in the social rewards system. For the system to work, participates must be selected using the weight of their known binary relationship with those they invite, and only the persons with a relational weight, calculated based on whatever known relationship strength indicators we have access to and can calculate, above a to-be-determined constant should be offered entry into the program. If both of the participants accept the terms of the program, they will receive rewards based on the activity of their partner, incentivizing familiar communication urging the other to be involved in the LinkedIn community. The networks of the users and communities they are a part of can be monitored over time and compared to the networks of other user invite pairs not qualified, qualified and not selected, and those who opted out of the program. Hopefully, this data will reveal that the idea & it's implementation was a sucess.

My idea stems from Sandy Pentland’s lecture on Social Physics: How Good Ideas Spread, one of the many great Talks at Google. I do not have the time or resources to do this alone, but if the right team comes together, we might be able to make it happen.

Authors

Selby Kendrick

http://selby.io

http://www.linkedin.com/pub/selby-kendrick/50/289/650/

Approximately 10 years of experience with Linux computing, building server stacks, creating software & webapps, and playing around with programmer logic.

Contribute

If you want to contribute I'd be more than happy to hear your ideas.

More research is required, and everything is open for debate.

I'd prefer for discussion to be done here, on GitHub. Please learn how to use git, or just register and create an issue above.