DOC: Clarify analysis methods

[HaoZeke]

The current documentation is hyperfocused on using ASV as a generational benchmarking tool, eliding (except for a dump of the help text) single commit usage like asv show and even things like asv compare / continuous. Additionally, user stories like using the profile data could also be improved.

i.e. to close this out, add documentation on:

• Single commit analysis
• Comparing generated output files
• Updating from v1 API JSON to v2
• Using profile data (there is some documentation from ENH: Return pstats.Stats for --profile #1253 but needs to be consolidated)