Milestone 3 PixelPerfect PowerIdea 1: Task Classification

From crowdresearch
Jump to: navigation, search

Task Classification

Introduction


Crowdsourced tasks differ considerably in various aspects - nature of task, skill and knowledge required, estimated time taken, complexity and various other factors. It would greatly help if we tagged each task, and grouped them together. Moreover, our dark horse idea attempts to standardize the price of tasks, for which classification is highly essential.

We also note that our classification system must be broad and flexible enough to accommodate any new tasks which come our way.

Design Aspects


Picture8.png
  • Tasks can be classified broadly based on the following aspects :
    1. Skillset - mathematical, logical, visual, programmatic, rote-memory, handwriting etc
    2. Nature of Task - information retrieval, social media, surveys, OCR, translation, reviewing etc.
    3. Complexity of Task - simple, complex
    4. Knowledge - Science, Philosophy, Psychology, Nature, Law, Regional and specialized knowledge etc
Therefore, we need to design our tree carefully to avoid overlaps and redundancies.
  • The ontology tree type classification of tasks draws inspiration from Wikipedia and its category system. General direction of design shall therefore merit a study of how Wikipedia classifies its pages.
  • Adding, removing and re-positioning categories in the ontology tree will be required from time to time. Moderators or high-level users shall be allowed to do this. The interface to edit the hierarchy must be simple to use.
  • Moderators and high-level users should have permissions to tag or retag existing tasks, to help complete them.

Benefits


  • HITs can be found in an easier manner through dashboard sorting.
  • Helps workers increase productivity by understanding task quicker and grouping similar tasks.
  • Helps requesters convey their goal easily.
  • Standardization of task prices requires some degree of classification.