Winter Milestone 5 @niranga
Introducing the workers skills to the Boomerang Ranking
Brief introduction of the system
Boomerang, the rating system that uses in Daemo decide what are tasks that workers going to get in their future and what type of workers, requesters going to get for their task. Requesters' ratings of workers of tasks are used to give early access to workers that requester rates highly. However, Boomerang still can recommend a worker to the requester who are not familiar with the posted task because Boomerang doesn't consider the workers skill level related to that particular task when recommend.
- Requester post two tasks to the platform (Image tagging, Translation)
- When the requester posts the tasks Boomerang recommend the higher rating workers first
- However, among those workers, there can be workers who are not familiar with those tasks because those workers previously have worked with the requester on some other type of tasks( Surveys, Bookkeeping, etc.)
- So, if the requester selects one of the higher rating worker not familiar with the task, there can be a quality issue. To mitigate the risk we proposed to introduce a task category filter option to Boomerang system.
How is the system solving critical problems
When a requester create a task he should choose a category relevant to that particular task. For example, if it's a image tagging task, task should go under 'Image tagging' category etc. So when the requester post this task to Crowd sourced market place, The system choose highly rated workers first and filter them according the category they have rated. Then Boomerang recommend the workers who has higher rating in that particular category. That way, the requester will get the most suitable higher rated workers first.
Also, after completing the task, the requester has to rate the worker and that ratings will be saved along with the task category.
So, when another requester post a similar task, the system looks for a suitable workers that has the highest rating and the category the worker most familiar with.