Interstellar turbulence plays an important role in sculpturing the interstellar medium (ISM), yet it is still not understood what are the most dominant turbulent drivers. The most common way of studying turbulence utilizes the spatial power spectrum (SPS) applied on the whole data set, which does not provide a way of tracing local spatial variations of turbulent properties caused by stellar feedback, gravity, or ram pressure. Recently, Szotkowski et al. (2019) developed a rolling SPS method which estimates the SPS slope locally, effectively mapping out variations in turbulent properties. We apply the rolling SPS on the neutral hydrogen (HI) images from the Galactic Arecibo L-Band Feed Array HI (GALFA-HI) survey, which offers both highspatial and velocity resolution over a field of view that is 13,000 square degrees large. We present preliminary results of how the SPS slope varies spatially for different velocity ranges, local HI vs high-velocity HI.