UJ students use tech to strike a blow against tender corruption

A group of four University of Johannesburg (UJ) students have struck a blow against corruption in the tender space with an app developed as part of the South African leg of global student software development challenge, the 2016 Microsoft Imagine Cup.

The four, all Honours students at the UJ Academy of Computer Science and Software Engineering (ACSSE), hail from South Africa, Zimbabwe and Nigeria.

According to team leader Chukwudi Obodoekwe, from Anambra State in Nigeria, the drive to design this system initially came from his uncle (who is also his guardian): “When he heard we are entering for the challenge, he said ‘Government in Africa is bad. You young people are bright. So find a real problem, like tender corruption, and fix it.’ It wasn’t easy, but he convinced us to go for a tender management development project”.

“We believe the diversity of the our team played a major role in the success of the project. We are four student from three different African countries but we found out that we faced the same problem in terms of tender corruption,” says Obodoekwe.

According to a press release sent to Memeburn, a key part of the team’s approach was integrating tender management processes in one system, from ‘cradle to grave’. Their ‘Kriterion’ project apparently spans the advertising of tenders, through to the evaluation of bids using the complex criteria in algorithms published by the South African Government, to the awarding of tenders and post-project or post-milestone delivery ratings.

Another important feature allows industry experts to submit estimates of what a realistic bid for a tender should amount to. In this way, tender bids upwards of R20-million for a project that industry experts estimate should cost no more than R10-million, for example, can be identified.

Lastly, the Kriterion project added a layer of ‘triple-blind’ functionality: the bidders on tenders, the industry experts and the Government officials are all completely anonymous to each other on the system, to reduce influencing of the tender award process.

The team used Microsoft Visual C# for the backend, Microsoft SQL for the database communication, Angular JS and HTML for the front end and SignalR for the real-time updates of a chat forum for each tender project. The mobile application was developed on the Android platform.

“We wanted to do something different and make a small impact in society. People talk about making changes, or just criticise the government, but you need to ask what role you play in making the country better. We believe that a change can start from individuals. That’s why we identified the need to address problems with tenders and procurement systems. If this is implemented we believe it can go a long way to curb tender corruption,” says Obodoekwe.

Also on 26 April, a project aimed at assisting fraud investigators won the Innovation category of the challenge. Rhyno Heydenrych, another Honours student at the ACSSE, specialises in digital forensics, where he analyses mobile phone and computer data related to commercial fraud investigations.

“Sometimes I go with the fraud investigators when they interview persons of interest. They interview both people who report possible fraud and people suspected of acting fraudulently. It struck me that there may be a way of recording the interviews and getting more insight into the answers than we currently do,” says Heydenrych.

Heydenrych’s project ‘VSA_Forensic’ records a human voice at high quality on a laptop, analyses the voice recording using a variety of algorithms he found in published academic research, and displays a user-friendly graph indicating tension and anxiety levels.

Polygraphs, or lie detectors, are highly complex to operate, he says. Meanwhile, voice stress analysis, which looks at emotion and stress in a voice, is not a new concept, and there is a lot of published research about it.

“Polygraph data or voice stress analysis is not accepted in a court of law as evidence. However, if a fraud investigator, or a policeman questioning a possibly drunk driver or a person conducting a job interview can have additional information indicating that the person they question is getting more stressed, that can assist in interviews.

“VSA_Forensic cannot be used by itself to determine if someone is telling the truth, it merely presents the indicators and the investigator then needs to interpret the data along with the context. However, the analysis can help an investigator to know they are on the right track with their questioning,” he says.

From here, the winning teams will compete in their respective categories at the regional semi-final in the MEA region during May, each for a chance to participate at the World Wide finals in Redmond in July.