methodology utilizes our previous work that extracted clues from a query Q and used these clues as attributes in a learned model that hypothesized from Q the requisite content Qx of the x-axis, the requisite content Qy of the y-axis, and the noun words that do not belong on any of these axes Qn.