WinterMilestone 3 Crayons ReputationIdea:Improving Boomerang by Domain-Clustering and Worker-Ability-Tests

From crowdresearch
Jump to: navigation, search

Reputation-related Idea

Improving Boomerang by Domain-Clustering and Worker-Ability-Tests

The idea is to improve the reputation system by introducing two new concepts:

  1. Domain-Clustering
  2. Worker-Ability-Tests


In the current system, position of the task of a given requester R on the task-feed of the worker W in determined by the rating given by worker W to requester R. Incase there are several requesters with same rating or requesters without any rating, their global average is taken into consideration. We believe that better results may be obtained if we replace global average with domain-cluster specific average.

Lets say a worker W performs majority of tasks of domain D(coding, design etc). Now to get the rating of an unrated requester R, we should take the rating average of all the workers of the domain D instead taking a global average. Rating received with such a setting will be more worker-specific and hence will improve the task-feed of the worker.

Domain based clustering of Workers.


We want the tasks posted by the requesters to reach the right audience. In the current setting, it is ensured that highly rated workers will be amongst the firsts to receive the task of the respective requester but we can further improve on that by adding some worker ability tests.

The plan is to have some category specific tests in the platform, completely optional, which the workers may take to improve their profile score.Once the worker pass the test of a particular category, say C, he/she will receive category C related tasks. Hence, task of a category C posted by a requester R will reach both the workers who are rated high by the requester and workers who have passed the C category test. This will also help the new workers to mark their presence in the platform.

Milestone Contributors

  • Ameen Mohammed Khan @ameenkhan07
  • Farheen Nilofer @farheen
  • Sachin Sharma @beingcooper
  • Sarah Masud @sarah
  • Shivam Rana @tminima