What makes a good project?
In general, a good project topic includes some combination of technical challenge, business opportunity, human interest, and opportunity for students to experiment with new technologies. We would suggest aiming for at least two of these properties, perhaps three, but all four in a single project would be excessive. It is worth scanning some examples of design briefs from previous years: 2018 list, 2017 list, 2016 list, 2015 list.
There should be a variety of technical challenges within each project, so that different members of the team can focus on different aspects. A challenge that is already a focus of intense commercial competition, such as "invent a machine learning algorithm to predict stock-market prices", will be beyond the ability of undergraduates, and will lead to disappointment because the result is certain to perform poorly by comparison to the state of the art as achieved by well-funded teams of PhDs.
A novel product or business model is more interesting than a "me-too" idea. Business viability is a relatively low priority - anything with a vaguely plausible market opportunity is fine. We understand that most technical start-ups pivot from their initial pitch anyway, so we don't expect computer science students to achieve anything beyond a concept that attracts popular attention.
Many students appreciate the chance to think how the world can be improved through technology. We are happy to see students address problems related to disability, inclusion, or economic and social inequality - either locally, nationally or globally (note that in some years, over 50% of our undergraduates come from outside the UK).
Almost all computer scientists are gadget freaks. They love to work with the latest hardware and systems, and those opportunities inspire some creative teams. In recent years, we have seen great projects using VR headsets, wearable devices, location technologies, microprojectors and so on. If you have a particular device you'd like to see the students use, the teams often work with technology loaned by clients. The one thing to keep in mind is that they do not have much time to acquaint themselves either with new languages or complex API stacks.
Things that have failed in the past
See the list of projects that were pitched to students, but for which we weren't able to put together a team because too few students were willing to work on them. It's not always easy to see why these turned out to be unappealing, so may be no clear patterns, but at least these give some idea of what to avoid.
After keeping that list for a few years, some emerging patterns would seem to be that: our students are not motivated by projects related to road transport; they are not inspired by wine and food (apart from chocolate); they may be worried by projects related to surgical procedures; and they are not interested in poetry or dance. Despite this, we are always happy to explore the limits of their interests, so don't be too discouraged from proposing projects in these areas - just make sure they are appealing, and be prepared for cancellation.