For multiple purposes this data is of utmost importance. Here just listing few possible use cases.
Any Gujarati font cannot be ideally called complete if it lacks support in rendering certain character sequence. In other words, all glyphs has to be supported. However for all its practical purposes providing complete coverage may not be primary goal of font designer/foundry. In such cases, those glyphs that occur more frequently then others need to be supported first. Statistical analysis can help us in deciding which are those scripts.
Even if a font with complete coverage exists, it would increase the file size comparatively which is not desirable scenario in many use cases e.g. web fonts. In fact, a web font generator that only adds those glyph which are actually used in the site would not just be a cool thing but a sought after requirement. Here also analysis of all glyphs in a corpus helps.
For multiple purposes this data is of utmost importance. Here just listing few possible use cases.