Ms Uduakobong Ekpenyong had never thought of having deep-fried ice-cream. Until a friend recommended it to her, that is.

This simple endorsement of a snack by a close peer was the inspiration behind Ekpenyong's PhD research on the use of social networks in energy-efficiency projects. 'I was told about something and I tried it, then I told someone else about it and they tried it. This process made me realise that consumers can spread the word about something like energy efficiency and thereby increase the number of people actually engaging in it.'

This Nigerian-born researcher, who was based in the Department of Electrical, Electronic and Computer Engineering at the University of Pretoria, wants to encourage people to adopt energy efficiency as a way of reducing demand on the power sector and increasing usage of renewable energy.

She is passionate about communicating her scientific work in order to help ordinary people, but has had trouble doing so with her previous projects. 'My master's degree was quite technical in the sense that I could not really explain what I was doing to people outside my field of research, namely electrical engineering. I wanted to do something I could explain to other people and for it to have relevance, not just to electrical engineering but also to other fields of study,' she says.

Ms Ekpenyong discovered her chance to do this in the large amount of data available about social networks. She made surprising findings during her investigation, which changed her idea of how social networks operate. She knew that the best way to spread the message of energy efficiency in a community was to give that message to the most influential people, which she assumed to be the people with the most connections in that particular network.

She formulated a mathematical model that would identify the most influential people in a social network. 'I found instead that people who may be connected to few people, but whose connections are connected to more people, tend to be the real influencers in the community,' she says.

These results enabled her to determine that the greater the influence a person has in a community, the greater their potential to effect indirect energy savings. In other words, finding the most influential people in a network could help save energy.

Ms Ekpenyong sees her research as contributing to the need to reduce demand on the electricity grid in South Africa. She foresees her research being applied in other areas, such as the commercial industry where advertisers can target the most influential people using her mathematical model, which, she says, can also be used in digital networks such as Twitter.

Her ground-breaking research has already gained considerable publicity for the University of Pretoria. 'I have presented my findings at three international conferences,' she says, 'and quite a number of people have read and referenced my publication.'

Although she is currently taking a much-deserved break from research, Ms Ekpenyong says she will explore different options for information propagation in the future.