How to Choose a Final Year Project You Can Actually Finish
A five-filter method for picking a final-year topic, and the four warning signs that a project is going to run out of time.
Five filters, applied in this order
- 1Does it produce a measurable result? If the answer is only "it works", the viva will be uncomfortable. Pick something you can put numbers to.
- 2Can you get every component within a week? Anything on a long lead time is a risk you carry for the whole project.
- 3Does it use what your syllabus taught? You have to defend the technology choice, and defending something you were never taught is avoidable difficulty.
- 4Can it be split across your group? Four students and one indivisible task means three people with nothing to say in the viva.
- 5Can you demonstrate it in five minutes without special conditions? If it needs sunlight, a specific network or a half-hour warm-up, plan for that now.
Narrow beats broad, every time
The strongest final-year projects do one thing properly. A system that measures a single quantity accurately, logs it, handles the failure cases and is documented cleanly will be assessed better than one that attempts five subsystems and demonstrates three of them badly.
Examiners are not impressed by breadth; they probe depth. The moment you have four subsystems, at least one is shallow, and that is the one you will be asked about.
A realistic hardware timeline
- 1Week 1 — finalise scope, draw the block diagram, list every component and order them.
- 2Week 2 — build and test each subsystem separately on a breadboard. Do not integrate yet.
- 3Week 3 — integrate, move to permanent assembly, and start capturing readings.
- 4Week 4 — measurement runs, failure-case testing and rework.
- 5Week 5 — documentation, presentation, and rehearsal of the demonstration.
- 6Week 6 — buffer. Something will need it. If nothing does, you finish early, which is a pleasant novelty.
Four signs your project is in trouble
- You are still ordering components in week three.
- Nobody in the group can draw the block diagram from memory.
- The project works only when one specific person sets it up.
- You have no readings written down anywhere with two weeks to go.
Use your guide properly
Your guide has seen the failure modes before and can tell you in ten minutes that a topic will not work. Take them the scope and the block diagram early, not a finished project in week five.
Take specific questions rather than open ones. "Is this topic okay?" gets a shrug; "we plan to measure X with sensor Y at this interval — is that resolution adequate?" gets useful guidance.
