This chapter explores in detail about the data-orientated techniques which produce the policy spaces. There are many different ways to combine or ‘reduce’ an array of individual estimates to the lines or dimensions which constitute a policy space. Different majorities emerging, depending on the order in which policy alternatives are voted on. Most policy dimensions, even if they are substantially independent of each other, are weakly related to some degree. Factor analysis and similar dimension-finding techniques can create policy spaces based on their own purely methodological assumptions, though the way we interpret them usually brings in theory. In general, right-angled policy spaces fit with our theoretical thinking and are easier to work with. Spatial representations of policy processes and outcomes are central to both theory and measurement in political science. Regression equations can in turn be used as concise theoretical statements providing both an explanation of some political process and a prediction about its outcomes.