SJR is a prestige metric inspired by the Google PageRank algorithm and the idea that not all citations are the same. With SJR, the subject field, quality and reputation of the journal have a direct effect on the value of a citation and the impact that journal makes.
This metric is based on citation data from the Scopus database. It is freely obtainable on the Web from the SJR Journal Rankings page.
Challenges with SJR include that it is very difficult to validate and a very complex calculation!
The SJR calculation is based on the Eigenvector Centrality theory of social networks and consists of an iterative process which takes "the weighted citations received in year X for documents published in the journal in years X-1, X-2, and X-3." (SJR Help, n.d.).
More in-depth information about the calculation of this metric is available in an overview page and in a paper presenting the detailed formula.