We propose a methodology to evaluate the impact of a Digital Library’s (DL) collection and the characteristics of its user community by an analysis of user retrieval patterns. Patterns of journal and document co-retrievals are reconstructed from DL server logs and used to generate proximity data for journals and documents, resulting in a weighted relation defined over the DL document collection represented by a network of document and journals. A measure of discrepancy between user-defined measures of document impact and the Journal Citation Record (JCR) Impact Factor (IF) published by the Institute for Scientific Information (ISI) is used to analyze characteristics of the DL user community. A preliminary analysis of the Los Alamos National Laboratory (LANL) Research Library (RL) server logs registered in 2001 demonstrates the potential of this approach.