Data conversions are not the most exciting types of projects, but I find them interesting because each is unique, and I enjoy the challenge of doing a superior job fitting old data into new systems. It’s normal to encounter un- or poorly-constrained text fields that have accumulated a great deal of cruft.
I dislike the attitude of merely dumping raw data into the new system and letting the user hold their noses at the digital sewage. My attitude is that the machine should conform to the needs of the user. Time spent viewing dirty string values and cleaning them up with some logic and a series of regular expressions turns digital sewage into useful relevant information, and much happier users.