Apple's total desktop/laptop share has also been dipping pretty sharply for the last three months. That should be taken with caution, obviously, because the stats often fluctuate so it may not mean a whole lot, but Apple just hit a four-year low (6.6%), which even in the staggered walk of usage stats may be a bad sign (although they hit a pretty good high in March, so who knows).
On the other hand, Linux has been steadily and consistently rising, and hit an all-time high in the desktop/laptop space in June (1.7%). That's still small potatos, and it may just be a sign of demographic reshaping from the contraction of the desktop market (i.e. many people are replacing computers with android tablets, and the population that continues to use desktops/laptops may be, on average, slightly more inclined to use Linux than the total population of computer/device users, etc.).
What's interesting about the Linux situation though, is that windows usage took a dip in late 2013, but recovered and has otherwise been more or less flat for several years (the usage today is very close to the usage two years ago in the measures I've seen). Which means that the net effect is that Linux's half percent may have come out of apple's market share, and not out of windows', which I never would have expected in a million years. See, e.g.
http://www.netmarketshare.com/operating-system-market-share.aspx?qprid=9&qpcustomb=0&qpsp=163&qpnp=24&qptimeframe=M (In August 2012 the stats were W 91.77 M 7.13 and L 1.1, in July 2014 they're W 91.68 M 6.64 and L 1.68
Obviously one would have drawn very different conclusions looking at the same data six or eight months ago, and the data in a few months may shw a complete reversal, so grain of salt, etc.