stats « Blog - Giant Goat Web Development

March 30, 2011

Real time analytics

Filed under: fun,tools — Tags: analytics, chartbeat, clicky, stats — Justin @ 9:01 am

This week and last, I had the great fortune (misfortune?) to get loads of traffic to one of my personal websites from websites as varied as NPR’s OnPoint to Der Spiegel. I’m in the middle of evaluating 2 real time analytics tools, Chartbeat and Clicky, but I’ll use screenshots from Chartbeat since it let’s me rewind and show traffic at a specific time (Clicky looks a little less frightening and is probably what we’ll install for our customers).

Both tools let you know in almost real time (there’s a one or two minute lag) who’s visiting your site and from where. Chartbeat sent me an email to alert me that I had a lot of traffic (it defaults to “Send me an alert when total visitors on any goes above monthly max”) and I logged in to see:

It turned out a lot of my traffic was coming from Slashdot, a relatively large technology news site (“News for nerds, stuff that matters”), and I was getting more traffic than the server could handle (my personal blog is hosted on a very weak machine).

I was able to:

Redirect the traffic to a server that could handle it
Jump in to a few online discussions of my site and participate
Brag to my office mates

It’s a pretty cool experience, I think we’re going to start moving customers over from Google Analytics to real time analytics.

Comments (0)

October 13, 2010

W[h]ither Flash

Filed under: Uncategorized — Tags: flash, stats, technology — Justin @ 4:28 pm

We do a lot of work in Flash here, it’s everywhere on the web. I think if I had my way I’d drop it from our list of core-competencies ¹.

Recently though, I found myself asking why, between the recent exploit² that will take 2 weeks to close (and probably necessitate yet another Firefox reboot) and my realizing how fiendishly easy it is to get around Flashblock in Firefox:example here (warning, many contain unicorns) ³.

The easy answer is that it’s available on over 96% of our visitors machines, (compared to ~60% for HTML5), but that can’t be the whole story; Java is installed on 94% of our visitors’ machines and nobody’s clamouring for that ⁴.

Just looking over the tone of the last few paragraphs, I realize that I’m pretty down on Flash, and I think for me that comes from the pain we’ve experienced in the development environment (I’ve had more neutral experiences with Flex). So I’m a little excited that Apple seems to have declared war on them (and remember that Apple is the most valuable technology company on earth now) and that when they can get enough in stock, iPads cannibalize sales from laptop PCs … by as much as 50%.

You can always punt these questions to the customer, and in this case they preferred Flash. That was easy enough but I wonder if there will soon be a day when offering flash solutions seems as out dated as a “Best viewed in IE 6″ button.

but then again, I’d probably prune the list down to just “passionate technology advocacy” if I didn’t have adult supervision ↩
Edit: it’s not a zero-day exploit, since that requires an exploit to appear the very day the developers know about it. ↩
I’m happy to note that it’s easy to completely disable Flash in Chrome, by visiting chrome://plugins ↩
I like to think that this is because early Java web apps were ugly, and we make a lot of our decisions on emotions, but it’s interesting to note both technologies are 15 years old, with both Java and F(uture Sp)lash launching in 1995 ↩

Comments (0)

August 30, 2010

What the web looks like

Filed under: Uncategorized — Tags: stats — Justin @ 7:24 am

The clever folks at nmap.org downloaded all the icons from the 288 000 most popular sites on the web and scaled them according to the amount of traffic they receive. So to a first order approximation, here’s what the internet looks like:

Comments (0)

August 10, 2010

Risk Management Statistics

Filed under: Uncategorized — Tags: risk management, stats — Justin @ 4:28 pm

Some of our less public facing products deal with managing risks overseas, so I was a little too excited to find out that the US State Department publishes a listing of causes of deaths of Americans overseas (Thanks: Matt).

Here’s the breakdown, I’ve shown figures with and without Afghanistan and Iraq since the risks in those countries is borne mostly by a very specific kind of traveler (and I’m more interested in relative risks, the risks in an active warzone are very different):

US deaths overseas 2002-2009

What’s interesting to me, is that despite all my cleverness making our software figure out what countries are risks for terrorism and epidemics, the real risk management questions I should be asking are:

Do you intend to drive?
Do you intend to swim or boat
and conceivably a few mental health and drug related questions

The only risk I correctly handled was homicide, still I intend to use this data to test my risk predictions for various countries. I might try to estimate a table of risks per capita by cross referencing deaths with the level of travelers in a specific country (If someone has already done that, please let me know, the UN stats are pricey).

Raw data below the fold:
(more…)

Comments (2)