Programming practices when receiving and manipulating received TCP/HTTP data?

Should data manipulation once data is returned either with TCP or HTTP be received as byte arrays or is it an O.K. practice to receive it as a string? I've been trying to find some professional projects on github to get my answer, but have had no luck. Some examples of HTTPClient from Microsoft on MSDN usually make use of the GetByteArrayAsync(website) method, instead of GetStringAsync(website). Is there any reason why they would use GetByteArrayAsync instead of GetStringAsync, which would make data manipulation much easier right off the bat? Are there any advantages to using GetByteArrayAsync first instead?