There are currently 1.6 billion mobile devices like cellphones and PDAs in the world today, and the numbers are rapidly growing. Improvements in computation power and storage on the devices has led towards support for many new applications. However, as the demand for data consumption by high bandwidth applications increases, the limited and shared wireless resource is likely to become a bottleneck in terms of its capacities and ubiquitous availability. Solutions like overlayed heterogeneous networks, intelligent network deployment, and spectrum management try to tackle certain issues to improve wireless resource utilization, but a big problem faced by each of these methods is the difficulty in collecting real-time data about the wireless environment. Large-scale continuous monitoring of the wireless medium is difficult because of the huge scale of deployment and the dynamic behavior of the medium. We propose a novel solution to real time wireless monitoring that can help each of these methods to work more efficiently. We do this by collection of feedback of the network state from mobile devices and on-board vehicular computers, followed by analysis of the data in an online as well as offline manner. In this paper, we present an application framework to enable the use of mobile devices as sensors, and outline a number of interesting applications that can be created on the framework. We identify suitable tools to be used on the client as well as the server side, data models for efficient storage and retrieval of sensor data, and wavelet based mining techniques to work on this data. We show that making such a framework is a very rich research area that brings together wireless communication, databases, and statistics, and contains many open problems.