some statistical analysis might be useful as well. e.g. create 10 random graphs, extract their key metrics (connectedness, average node degree, ...), average them over the number of graphs, and check whether they are within some acceptable range..
this could help to spot problems with graph generation, which are technically valid, but not intended. e.g. a graph with no relations (only nodes) would always be a valid graph. but it's unlikely to be a useful graph
some statistical analysis might be useful as well. e.g. create 10 random graphs, extract their key metrics (connectedness, average node degree, ...), average them over the number of graphs, and check whether they are within some acceptable range..
this could help to spot problems with graph generation, which are technically valid, but not intended. e.g. a graph with no relations (only nodes) would always be a valid graph. but it's unlikely to be a useful graph