If you're interested in research, pick an interesting subject, and look for an open research question/projects in new published articles.
If you're interested in impact, i think it's easier to build an influential project(influencing 1000+ or more people in a reasonably important way) than a meaningful research. There are many useful research papers by academia that haven't been developed or commercialized and could be the base for useful projects. For example: [1].
[1]ParkNet: Drive-by Sensing of Road-Side Parking Statistics
If you're interested in impact, i think it's easier to build an influential project(influencing 1000+ or more people in a reasonably important way) than a meaningful research. There are many useful research papers by academia that haven't been developed or commercialized and could be the base for useful projects. For example: [1].
[1]ParkNet: Drive-by Sensing of Road-Side Parking Statistics