You can't. That's the magic of data abstraction.
http://en.wikipedia.org/wiki/Abstraction_%28computer_science%29
Binary is simply the electrical signals your hardware deals with. Asking to "see" the binary is basically asking to see each individual electrical signal.
With some low-level programming or IT skills, you can see representations of things in binary, but those representations themselves aren't the binary. It's like a picture of a picture. They're not the actual binary.
http://en.wikipedia.org/wiki/Binary_file
For a piece of software:
Binary -> Machine Code -> Assembly -> Language(s) -> Programs
Languages can include C, C++, C# Java, etc, or any combination of languages.
For data:
Binary -> Electrical signals
0 = low electric signal.
1 = high signal.
CPU interprets the data based on how the program/OS it is running tells it to.
For data storage, it depends on the medium, but the idea is the same: have one thing represent zeroes, and another thing represent ones. Store binary data in the medium. This can be anything from:
- Punch cards
http://en.wikipedia.org/wiki/Punched_card
- Magnetic tape
http://en.wikipedia.org/wiki/Tape_drive
- Magnetic hard drive
http://en.wikipedia.org/wiki/Hard_drive
- CD/DVD/BluRay
http://en.wikipedia.org/wiki/CD-ROM
- NAND flash devices
http://en.wikipedia.org/wiki/USB_flash_drive
http://en.wikipedia.org/wiki/Solid-state_drive