Reconstruction is a facility for distilling diverse qualitative data within consolidated texts. This is especially helpful for the management of data which represents hidden and counter cultures and has a panoramic complexity and emergent realities that are difficult to capture in more established forms of presentation. Reconstruction helps to represent the original coherence of the data while showing the interconnectedness necessary for thick description, and addressing the difficult interplay between the principles of submission, emergence and the personal knowledge of the researcher. This paper explores how this might be done in the case of visual and descriptive data. In this endeavour, lessons can be learnt from alternative attitudes to research and data in mainstream education and fine art.