MobileActive.org

During a study conducted in January 2010 by researchers of the Columbia University Mechanical Engineering Department in New York City, approximately 300 farmers were surveyed in rural Mali. Farmers were randomly sampled via standard proportional, stratified, cluster techniques. Data collection took place through the use of HTC G1 smartphones running Google’s Android operating system. The phones were equipped with Open Data Kit (ODK) software; a system that immediately digitizes data for analysis, allows for remote monitoring of the collection progress, and facilitates the gathering of data, eliminating the need for paper surveys and therefore significantly reducing survey times. ODK has the potential for a profound impact on the future of data gathering, particularly in development applications where locations may be remote and budgets tight, yet where mobile phone use is rapidly increasing with the expansion of service
coverage.

Over four and a half billion people live in the developing world and require access to services in the financial, agricultural, business, government and healthcare sectors. Due to constraints of the existing infrastructure (power, communications, etc), it is often difficult to deliver these services to remote areas in a timely and efficient manner.

The CAM framework has found success as a flexible platform for quickly developing and deploying high-impact applications for these environments. Many of the applications built with CAM have relied on a model where a field worker with a mobile phone regularly returns from a disconnected environment to one with connectivity. In this connected state, the phone and a centralized server can exchange information and get the collected data backed up on reliable media.

We propose extending CAM’s networking model to enable continual operation in disconnected environments. Using a set of heterogeneous paths made available through social and geographic relationships naturally present among workers, we describe a system for asynchronously routing data in a best-effort manner.