Because I suspect that presenting the data this way masks the reality of why there are sheer number of people who can't find work. But even without major (which I was simply curious about), the data set is suspiciously aggregated.
For example, what if we were to drill down into the data and discover that in the metric "Some college or degree" that 99% of the individuals who could not find work were in the "some college" category and only 1% were in the "degree" category. That would paint an entirely different picture: having a college degree is incredibly useful, but attending college without completing the degree is not very useful in terms of being employable.