Certainly the type of content that takes up the most space would be uncompressed raster images. Text and vector data - in general - take up very little space. However, there are caveats to all of that. Is compression being used for the data (any type) and if so, what kind? Lossy image compression (eg. JPEG) is good for some images but really bad for others. Are the fonts associated with the text embedded? And if so, are t subset or full embedded? What colorspace(s) are being used? CMYK and Spot color data (any object type) take up more space than RGB or Grayscale data. PDF also supports audio, video and 3D data - all of which can potentially be large when fully embedded or not compressed properly. And then there are all the non-visual things that can be present in a PDF that would take up space Metadata Structure & Semantic information Navigation aids etc. But at the end of the day - none of this is PDF specific. These are the same questions you would ask for Word or Keynote or EPUB or any other format. PDF just has lots more options for both increasing and decreasing file size.
PDF's main advantage is that it's open source, which means that there's a plethora of quality standards.