Let's assume you are working for the NSA and wanted to use Python to profile large amounts of e-mail that you are intercepting.
Write three fictitious e-mails, one a business e-mail, one a love/romance e-mail and one e-mail from Ladun Bing Osman to an accomplice that threatens your country.
Now use Python's text handling prowess to profile the three letters, so the dastardly letter can be identified. Check if fragments of words can do this.
Note: As of this time there are over 10 billion (yes, billion!) e-mails each day worldwide. Would be tough to read them all in person! To this you can add about 15 billion spam e-mails!