Hacker News Reader

Comment by Morendil | Hacker News Reader

(This is a time series graph of the count of my GMail spam folder since 2009, updated daily.)

That's almost bump-for-bump identical to my data, only I show data going back to 2002, and you're only showing back to 2009. The correspondence is remarkable:

http://www.solipsys.co.uk/new/SpamGraphs.html?HN

That's one reason I set up this cron in the first place - I wanted to see if "my spam" was the same as "everyone else's spam".

I wish I'd had your foresight and set this up earlier, but now I'm glad I didn't let that prevent me from doing it anyway; it's always fun to play scientist. :)

??? You are getting more than 4,000 spam emails a day now and were above 30,000? Do you have your email listed publicly?

Also how do you automate the counting and storing it in the spreadsheet? I think that's pretty damn cool.

It's per-month, not per-day - GMail only stores 30 days' worth of email in the Spam folder.

As to how - cron job on my server querying GMail via IMAP. Source below. Manually copy and paste the resulting text file to a GDoc spreadsheet; use the Time Series widget to generate the chart.

    require 'net/imap'
    require 'date'
    imap = Net::IMAP.new('imap.gmail.com',993,true)
    imap.login('yourlogin','yourpwd')
    msgs = imap.status("[Gmail]/Spam",["MESSAGES"])["MESSAGES"]
    puts Date.today.to_s + "\t" + msgs.to_s

Does this graph show some seasonal pattern?

Doesn't look like it, but perhaps a closer analysis would show some. One thing it shows for sure is a change, early on, in GMail's algorithm for deleting old mails left in the Spam folder - that's where it goes from jagged to smoother. And it confirms the article's "spam at its lowest" story.

Reply on news.ycombinator.com