by taking statspack snapshots over time, it'll gather much more than you ask for - but you will likely ask for more over time.

short of that, you would set up a job to capture the number of user commit (v$sysstat) and a timestamp and v$instance.startup_time. This job would run at least daily (so you can see the daily increase). But bear in mind - that you will lose some counts across database "shutdowns" (unless you incorporate into your shutdown scripts a final "count the number of transactions").

ops$tkyte%ORA9IR2> create table t ( tim timestamp, dt date, commits number );

Table created.

ops$tkyte%ORA9IR2>
ops$tkyte%ORA9IR2> insert into t
  2  select systimestamp,
  3         v$instance.startup_time,
  4             v$sysstat.value
  5    from v$instance, v$sysstat
  6   where v$sysstat.name = 'user commits'
  7  /

1 row created.

ops$tkyte%ORA9IR2> begin
  2          for x in 1 .. 100
  3          loop
  4                  insert into t2 values (x);
  5                  commit;
  6          end loop;
  7  end;
  8  /

PL/SQL procedure successfully completed.

ops$tkyte%ORA9IR2> insert into t
  2  select systimestamp,
  3         v$instance.startup_time,
  4             v$sysstat.value
  5    from v$instance, v$sysstat
  6   where v$sysstat.name = 'user commits'
  7  /

1 row created.

ops$tkyte%ORA9IR2> select *
  2    from (
  3  select tim,
  4         dt,
  5             commits,
  6             lead(commits) over (partition by dt order by tim) next_val,
  7             lead(commits) over (partition by dt order by tim) - commits  diff
  8  from t
  9       )
 10   where next_val is not null
 11  /

TIM                            DT           COMMITS   NEXT_VAL       DIFF
------------------------------ --------- ---------- ---------- ----------
01-FEB-07 02.42.49.703262 PM   31-JAN-07        323        423        100

you would then be able to trunc the dates down to whatever you wanted - aggregate the counts and using lead - produce the report.