Networks are developing very quickly in the social sciences, and they are beginning to emerge in art history. This paper explores the making of network visualizations, from the building of the dataset to the analysis of results. Starting from an initial corpus regarding the Parisian auction sales for modern paintings, we developed a methodology to enrich it, thanks to linked open data platforms and technologies for realigning datasets. We then call into question the visualization of networks. Although it brings about an overview of the market and allows a very close reading, the best is the enemy of the good: too much information and too many nodes may represent a handicap, hence the need for both cutting networks and combining them with other modes of visualization.