Using a metric related to the returns correlation, a method is proposed to reconstruct na economic space from the market data. A reduced subspace, associated to the systematic structure ofthe market, is identi0ed and its dimension related to the number ofterms in factor models. Examples were worked out involving sets of companies from the DJIA and S&P500 indexes. Having a metric de0ned in the space ofcompanies, network topology coe8cients may be used to extract further information from the data. A notion of “continuous clustering” is de0ned and empirically related to the occurrence ofmarket shocks.