Observation or problem
There is a real issue worth paying attention to.
Research Training
Engineering Project or Research Project?
Building something useful is valuable. A Master's project must do one more thing: it must answer a research question using evidence.
The Main Shift
A practical problem is the start. A solution is the engineering response. Research begins when you ask a question, define evidence, compare against a baseline, and produce knowledge that other people can judge.
The Simple Difference
Use this table when you are revising your proposal. If your draft sits mostly on the left, your next job is to move it to the right.
| Engineering Project | Research Project |
|---|---|
| Builds a solution. | Answers a question. |
| Starts with a problem. | Starts with a problem and a knowledge gap. |
| Produces a product, tool, model, app, dashboard, device, controller, or simulation. | Produces evidence, analysis, and a contribution that others can judge. |
| Success means the thing works. | Success means we learned something defensible. |
| Evaluation may be basic testing. | Evaluation compares against data, a baseline, a method, or theory. |
| Reuse is mainly practical. | Reuse includes knowledge, a method, dataset, framework, benchmark, design principle, or validated insight. |
The Bridge From Building To Research
Use the scientific method as a thinking tool. The build is allowed, but the build must create evidence.
Engineering response
I can build a solution, prototype, dashboard, model, app, simulation, controller, or device.
Research question
What exactly do I want to find out?
Variables and measures
What will change, and what will I measure?
Method
How will I collect, generate, or analyse evidence?
Baseline
What will I compare against?
Analysis
How will I judge whether the result is meaningful?
Contribution
What can another researcher reuse, test, challenge, cite, or build on?
AI App
Not yet research: I want to build an AI app that gives useful advice.
Research framing: How accurate, explainable, and trustworthy are AI-generated recommendations compared with a rule-based or expert-reviewed baseline?
- Define the dataset boundary.
- Define the baseline comparison.
- Measure accuracy, usefulness, explainability, and trust.
- State the privacy, ethics, and governance boundary.
Digital Twin
Not yet research: I want to make a simulation of a system.
Research framing: Can a sensor-informed digital twin predict faults, energy use, or maintenance needs better than a simple model or existing monitoring approach?
A digital twin is not just a 3D model or simulation. It normally needs a real system, data connection or calibration, and a purpose for prediction, monitoring, optimisation, or decision support.
Case Study
Not yet research: I will build a dashboard for one facility.
Research framing: What design method, model, or evaluation framework can be learned from this facility and reused in similar facilities?
A case study can be research, but only if it produces knowledge beyond the single case.
The Master's Scope Test
- Can I finish this with the time, tools, data, and supervision available?
- Am I trying to build a whole company product instead of testing one research question?
- Do I need data I do not yet have permission to use?
- Am I creating a dataset, platform, model, and evaluation framework all at once?
- Is this becoming a PhD-sized project?
Before Your Next Supervisor Meeting
- My practical problem is...
- My research question is...
- The gap I think exists is...
- The system, data, site, or simulation I can access is...
- The variables or outcomes I can measure are...
- My baseline comparison is...
- My expected contribution is...
- The main risk is...
- The part I need help narrowing is...
This Is Normal
If your idea is currently mostly an engineering project, that is not a failure. Most good research starts with a practical problem. The next step is to turn the useful thing you want to build into a question that can be tested with evidence.