The U.S. baby name data comes from Social Security records available online at http://www.ssa.gov/OACT/babynames/.
It contains the national popularity of each of the top 1000 male and female names for each year since 1880. It also contains state-specific popularity of the top 100 male and female names for each year since 1960.
State-level popularity data is only available for the top 100 names in each state.
If hovering your mouse pointer over a state reveals that the selected name is not in the top 100 for that state for that year, the map is working properly.
The site uses the raw data from the SSA (see above). This allows you to see the popularity of name variations over time. When you search for a specific name, it lists similar names, allowing you to quickly compare to other variations with a quick click.
This site uses the raw data from the SSA (see above). If you're concerned about the quality of the data, please peruse their warnings at http://www.ssa.gov/OACT/babynames/background.html.
The popular in one decade tables list names that were most unusually popular in each decade, not the absolute most popular name (which is often a classic name popular for many decades). Michael was the #1 name in the 70s, just like it was from 1954-1998. Jason, on the other hand, had a crazy peak in the 70s, surrounded by steep dropoffs in the 60s and 80s (check it out). Being named Jason suggests having been born in or very near the 70s, whereas being named Michael does not.