The most common reason students give for not starting research is access. No lab, no mentor, no equipment, no funding. It is a real constraint, and it is also less limiting than it seems.
Start with a dataset, not a topic
Broad topics stall. Datasets create momentum. NASA, NOAA, the World Bank, the USGS, and dozens of national agencies publish enormous archives for free. Browsing what exists is often the fastest way to find a question worth asking, because you can see immediately what is answerable.
A good first project answers a small, checkable question about a real dataset. Not "how does climate change affect agriculture," but "how has growing season length shifted in this county over the last thirty years."
Write the method down before you run it
Writing your plan first is the cheapest way to catch flawed reasoning. It also makes your work reproducible, which is what separates a project from a chart.
Publish something, even if it is small
A clean writeup posted publicly does more for your development than a perfect project that never leaves your laptop. People respond to work they can read. Feedback from strangers on a modest analysis is worth more than praise on an idea.
This is the premise behind Rise2Research. The talent is not concentrated in schools with laboratories. The infrastructure is, and infrastructure can be rebuilt.