In this paper it is shown how new visualization techniques are being used to analyze the first results of the British 1991 Census and other large data sets. Of interest here are questions about how localities have developed over time; which neighbourhoods have experienced gentrification and in which places have the recessions of the previous decades had their worst effects? Overall, do we see a picture of polarization or a levelling out of social disparities either locally or nationally? It is argued that these questions cannot be answered by conventional quantitative techniques because the answers are unlikely to be simple enough to be presentable in tables or by equations. Pictures are needed to show how different processes occur in different places, and holistic patterns need also to be seen without generalizing out the detail. Neither traditional thematic mapping nor commercial geographic information systems can do this well. Spatial visualization is an alternative approach in which the researchers choose what they wish to see and how they wish to view it. Many problems require new methods of visualization for their exploration. A new census presents us not only with new statistics, but also with the opportunity and impetus to develop radically different ways of envisioning information to reveal more fully the human facts contained within a mass of social statistics.