Inditations map

I think the approach is the same for any 3D program.
I would start this project in Adobe Illustrator or Affinity designer and create the map and associated curves.
I would export svg of various levels of the drawing to create surfaces from. Ground plane, parking lot plane, grass, plane, walls, and floors.

I would then extrude the SVGs as needed and the fill-in faces where appropriate. I would export the Illustrator or Designer file as a PNG or JPG and use it for a texture map, I would also include in the map the up and down states as required.

I would then UV map all the faces to the map. Test render, make sure its looking good. Go back and setup up some camera constraints.

Then I would create a camera and lots of camera targets for the various locations you want the camera to zoom to.

I would then proceed to optimize the objects triangle count.

At the end I might have just one object and 1 material. I might consider three materials and three objects (base objects with btn hover, btn down objects, btn up objects). I don’t think this is very heavy so optimization may not be that important.

I would setup up HDR lighting but I would set it up so the background would be invisible.

I would then export to GLTF and get it into V3D.
You would create an app with a white background.

Get your puzzle work done for onclick, on hover.
Setup your cameras to tween to the correct location on click, making sure to leave room in the upper right corner for some HTML overlays.
Annotations, my personal choice would to do them in HTML/CSS
You will want to send an event to the parent window on hover to change the state of the annotation to get the annotation to change name as it hovers over the correct object. On click will trigger another event that will tell the parent window to open a DIV with the HTML stuff you want to show.

Thats about it.

There are a number of ways to do this. Hopefully, you are going to get paid for doing this.

Given there are three levels and three points of moving up and down in the building and lots of stores there are lots of methods to get from point A to point B.

There is Pathfinding Plugin …that if it could be modified might be useful and save a lot of work. Otherwise…brut force.

I would create a table of paths. As an example lets say there are two floors and four stores (A, B, C and D). A & B are on the first floor and C & D are on the second floor.

Starting with the second Floor…you have to go through the escalator to change floors. So for Store C and D, I would create one animation from each store to the escalator and one from C to D.
For the first Floor, I would create an animation to the escalator for each store and one for connecting A and B.

I would then create a look-up table. First Field Is the starting location and the second field is the destination. Knowing these and using the look-up table I know the route and can trigger the necessary animations.

The more stores, levels and up down points, just adds complexity to the problem but its the same problem.

You could also substitute the animations for just empty locations endpoints of the animation and dynamically draw something between the points as an option if you didnt want to create that many animated assets.