Make sense of *BIG* Web data is facing 4Vs problems: big Volume, high Velocity, high Variety, and unknown Veracity. We propose to build a virtual layer of Information Sensor on the Web. An Informaton Sensor is a programmable “focused crawler” that continuously discover, extract and aggregate structured information around a topic.