English phonology, American English dialects’ (and other dialects’) /r/ is usually pronounced retracted, post-alveolar/pre-palatal (usually bunched/molar), transcribed something like [ɹ̠ᶹ], so it causes alveolar consonants in the same cluster to retract/palatalize, usually into a post-alveolar affricate ([d͡ʒ] – the “j” sound for voiced stop /d/, [t͡ʃ] – the “ch” sound for voiceless stop /t/, [ʃ] – the “sh” sound for voiceless fricative /s/). The term would be assimilation (of place of articulation).

“Dragon” /dræ.gən/ -> [dɹ̠æ.ɡɪ̈n] -> [d̠ʒɹ̠æ.ɡ(ɪ̈)n]

You can see the same thing with words like “tree” /tri/ -> [t̠ʃɹ̠i] or even “street” /strit/ -> [ʃt̠ɹ̠it]

Would explain simpler but can’t, break ends now, just know its because consonant pronounced in different place in mouth is conforming to being pronounced in the same place in mouth as other consonant that is right beside it (like with “in-” vs “im-”, “impractical”, which notably isn’t “inpractical”, or “incandescent” which notably isn’t “imcandascent”, or “indecisive” etc. etc.)