Soluble and insoluble impurities play a crucial role regarding the deformability and thus the flow of polar ice. To better understand this interplay from a mechanistic point of view it is especially important to investigate the location and chemical composition of micro-inclusions (Stoll et al., 2021), which are among the most abundant impurities in polar ice. New results from a systematic analysis of micro-inclusions in Holocene ice from the East Greenland Ice Core Project (EGRIP), which has been drilled near the onset of the Northeast Greenland Ice Stream (NEGIS), offer unique insights into the dynamics of fast flowing ice over different scales, ranging from kilometres to micrometres. Investigating the small-scale properties of eleven samples from Holocene ice, i.e. the upper 1340 m of the EGRIP ice core, we mapped the locations of several thousand micro-inclusions inside the ice. The use of cryo-Raman spectroscopy allowed us to obtain a representative overview of the mineralogy of these inclusions in the ice without the risk of contamination. We identified a variety of Raman spectra, mainly from sulphates (dominated by gypsum) and terrestrial dust, such as quartz, mica and feldspar. The observed mineralogy changes with depth and EGRIP Holocene ice can be categorised in two different depth regimes, i.e. the upper (100-900 m) and lower (900-1340 m) regimes, depending on their mineralogy. Furthermore, micro- inclusions show certain spatial patterns, such as clustering or layering, which are partly related to their mineralogy. We thus conclude that Greenlandic Holocene ice has a broader, and more variable, mineralogy than previously reported and that chemical reactions might take place within the ice sheet, possibly altering the paleo-climate record. Our approach further demonstrates the added value of systematic, combined high-resolution impurity and microstructural studies, and the importance of considering different spatial scales and is thus another step towards a more holistic understanding of impurities in ice.