Statistics of the largest eigenvalues and singular values of low-rank random matrices with non-negative entries

We compute analytically the distribution and moments of the largest eigenvalues/singular values and resolvent statistics for random matrices with (i) non-negative entries, (ii) small rank, and (iii) prescribed sums of rows/columns. Applications are discussed in the context of Mean First Passage Time of random walkers on networks, and the calculation of network "influence" metrics. The analytical results are corroborated by numerical simulations.