Data Viz KitRights  |  Data  |  Questions  |  Charts  |  Hazards  |  Improvements

Crowdsourced Data

The term “crowdsourced data” refers to two main types of data: data collected and submitted from a wide variety of sources, and data analyzed or processed by a large group of “microtaskers.” Microtasking is where large tasks are split into smaller tasks that individuals can complete over the internet with little or no training. Microtasking is especially useful for tasks that are repetitive, but require human judgment and cannot be completed accurately by software, such as categorizing content, audio transcription, or identifying features in an image. Amnesty International used its extensive membership network to help examine thousands of satellite images of remote parts of Darfur where bombings and chemical weapons attacks are suspected to have taken place. Crowd-sourcing enables large amounts of data to be collected and processed with great speed, and at a relatively low cost. Volunteer-contributed data is valuable when information about human rights violations is dispersed and not easily gathered. For example, Ushahidi was developed in Kenya to allow victims of electoral violence to submit data about their experience to a single online hub. While such data can be crucial in emergency or otherwise insecure settings, it may also be prone to error, or skewed by accessibility. Crowdsourced data can sometimes be processed using statistical tools to improve the accuracy and quality of the data collected.