5

I'm working on a problem merging some internal customer data on email address. However, I know that different email address can resolve to the same person. Some examples are:

  • mapping a vanity domain to your free web mail (e.g. example@gmail.com and example@example.com)
  • company mergers like BigCorp.com and HugeCorp.com merge, creating BigHugeCorp.com (so first.last@bigcorp.com and first.last@bighugecorp.com are the same person).

In these case, emails are delivered to the same inbox. I would like to merge these, so I am looking for an open database of email domain aliases. I'm not even sure domain alias is the right term, but Google Apps uses this in Give users an address at another domain.

Potentially this could leverage answers from Dataset of domain names or Open database of domain registration information? if you could query on matching MX records.

Community
  • 1
Kevin Hakanson
  • 151
  • 2

1 Answers1

1

This is a tough question because email addresses are not openly available, so getting a dataset of mapping between addresses is even harder.

I suggest to take a big email corpus like the Enron one, and then build a tool that follows email threads and maps FROM and TO in subsequent emails.

Email in MBOX format is well enough structured to mostly read only the header information, and probably not need to parse the text

Since the Enron Corpus is grouped by person, you can find all the aliases that person uses:

http://www.enron-mail.com/email/kuykendall-t

You can also read the header and body text from the emails themselves:

http://www.enron-mail.com/email/kuykendall-t/discussion_threads/FW_Fw_sick_but_funny_jokes_1.html

The email header FROM states:

From: tori.kuykendall@enron.com

The body of this email states

Forwarded by Tori Kuykendall/HOU/ECT on 10/31/2000 07:29 AM

And the previous email was written to

To: tkuyken@enron.com

Presumably tkuyken@enron.com and tori.kuykendall@enron.com are both Tori Kuykendall, corresponding to the folder

http://www.enron-mail.com/email/kuykendall-t

If you search Enron on here, you'll find lots of tools, various export formats, etc

https://opendata.stackexchange.com/search?q=enron

philshem
  • 17,647
  • 7
  • 68
  • 170