Imperfections, often called ‘uncertainties’, exist in almost every spatio-temporal dataset, especially in historical data. They are of different types (unreliability, inaccuracy…) and concern every data dimension (space, time and theme). Based on previous work, this article proposes a synthesis qualitative classification of imperfection types. This classification has been assessed with domain experts (hydrologists, geophysicians and GIScientists working in a railway company) during an experiment, that gave positive results towards the use of this classification. Participants were also asked to evaluate the seriousness of each imperfection type in an analysis context. This evaluation has allowed to associate a quantitative index to each imperfection type and to visualize a quantity of imperfection attached to each spatial object in a map.