We propose techniques to evaluate and optimise the performance of humanitarian organizations at a country scale using open online data. To evaluate performance, initially three data based indicators are proposed, which measure: how humanitarian organizations in a country supply aid proportionally to the demands; how able is an organization to reach the provinces more in need; and how well is aid being coordinated by regions of neighbouring provinces. Then, we propose other four indicators based mainly on network centrality measures, which are calculated from a network of partnerships between organizations, that for many countries can be inferred from a popular type of dataset. These four indicators measure: how big organizations are taking advantage of their size to make partnerships and influence others; how influential organizations tend to have partnerships among them and less with not so influential organizations; the role of international organizations as connectors between organizations; and the participation of organizations in doing partnerships. To optimise performance, we propose an algorithm that redistributes aid geographically in such a way that optimises some of the data and network based indicators proposed. The results were promising when we tested our ideas in the Afghanistan humanitarian scenario of 2015. We also compared some indicators calculated for Afghanistan with other seven countries, and there were good insights obtained.