sum case를 사용했습니다.

with t as (
select 1 매체, 'a' 구분, 15 값 from dual
union all select 1 매체, 'b' 구분, 20 값 from dual
union all select 1 매체, 'c' 구분, 10 값 from dual
union all select 2 매체, 'a' 구분,  4 값 from dual
union all select 2 매체, 'b' 구분,  5 값 from dual
union all select 2 매체, 'c' 구분,  6 값 from dual
)
select 구분, 
       sum(case when 매체 = 1 then 값 else 0 end) col1,  
       sum(case when 매체 = 2 then 값 else 0 end) col2
from t
group by 구분

pivot 함수는 이렇게 쓰면 되는군요

with t as (
select 1 매체, 'a' 구분, 15 값 from dual
union all select 1 매체, 'b' 구분, 20 값 from dual
union all select 1 매체, 'c' 구분, 10 값 from dual
union all select 2 매체, 'a' 구분,  4 값 from dual
union all select 2 매체, 'b' 구분,  5 값 from dual
union all select 2 매체, 'c' 구분,  6 값 from dual
)
select *
 from t
pivot (sum(값) for 매체 in (1 as col1, 2 as col2)
);