17 comments

[ 15.0 ms ] story [ 1904 ms ] thread
There is no way I will let some third-party online tool access my inbox. Privacy and security are much more important to me than some fancy gimmick visualisation.

I would love to do this kind of analysis on my mails. Any chance for a local offline tool?

I wrote software which scans a hard disk for email in any of 6 formats, converts them to Mbox (eg, standard, with emails one after another in plaintext), creates 2 indexes (full-text and SQL) then lets you search all your emails with a Lucene query (full text of email bodies) or an SQL query (MySql or PostGres or SQLite - SQL database of common email headers). It doesn't run on 'the cloud'; it's a Windows application.

An example search is 'All emails from "geocities.com" between March 2009 and April 2009, with attachments.

I haven't found anyone who wants such a program. Anyone here interested?

Absolutely. It fits my usage nicely.
If you're interested, we could talk over email.

Do you have an email address or do you want me to post my email address? Or not? ;-)

I updated my profile to show it, I think.
Why mbox instead of, say, Maildir? The irreversible mangling mbox requires makes me reluctant to use it as an archive format.
It can output single emails to folders named after months (eg, 2011_10_18), which I think is MailDir.
I'm interested too... Jeremy@Zawodny.com (since I've considered writing such a beast myself like a dozen times)
For Linux users there is "notmuch" for full-text searching. http://notmuchmail.org

Uses the excellent Xapian library to perform just about any search you'd need.

The article says the tool runs locally.
MUSE is a Java applet that runs locally.
If you click on "Other accounts" at the top then you can point it at an mbox folder, if that's what you mean by "local offline".
Does not seem to like Gapps.

Great application of the archive concept though.

I'd love something similar for IRC logs. I know there are all those joke irc log analysers, but they seem to be mainly for channel logs.
I wrote 'mu' (http://www.djcbsoftware.nl/code/mu), which let's you index/search maildirs; I'm currently adding glue to use it as an emacs-based e-mail client as well (basically, something between wanderlust and notmuch).

Anyway, to get back on-topic, as an alternative way to analyse your mails, I've added guile (scheme) bindings, and it's pretty easy to get all kinds of interesting statistics from your mails -- see https://gitorious.org/mu/mu/blobs/master/toys/muile/README

> It’s sad but inescapably true: Your email inbox is your life.

I'm a bit surprised by the nostalgic tone of the first paragraph. Why is that supposed to be sad? I find it pretty cool, because it is so easy to keep an archive of every mail you send or receive, unlike paper mail.