• find . look in current directory
• -type f -empty -prune -o -type f all non-empty files
• -printf "%s\t" -exec file --brief --mime-type '{}' \; print file size, [tab], and run the file command to obtain mime-type
• | awk 'BEGIN {printf("%12s\t%12s\n","bytes","type")} pipe to awk, print column headers
• {type = $2; a[type] += $1} build a dictionary of mime-types, increment value by file size. this creates a running total of size per file-type.
• END {for (i in a) printf("%12u\t%12s\n", a[i], i) |"sort -nr"}' print the result, sorted

Example output:

       bytes            type
    10422306    application/x-gettext-translation
     6759737      text/plain
     6653499      text/x-php
     2316621    application/octet-stream
     1486637       text/html
      714594    application/zlib
      524756       image/png
      210196    text/x-pascal
       45489      text/x-asm