Much of the data you need to collect can be automatically gathered by System Center. But you will need a lot of information.
* All servers running WS2003
* All non-WS2003 servers that touch the old servers in one way or another.
* Network and storage connections to 2003.
* All of the apps running on them. Then you need to check if those apps have new, current versions or if they are abandoned or no longer supported and make contingency plans for that.
* All data stored on them.
* Who accesses the WS2003 servers via Active Directory.