Skip to content


Data Porn – Express Edition

I’ve realized recently that I like data.

 

I realllly like data.

 

I blame quantitative research and machine learning (and Uiharu).

Uiharuuuu

Yeah, you heard me.

 

To be more specific, I’m starting to enjoy collecting data and trying to see if anything cool comes out of it. Professionally, this usually involves analyzing software and software communities. If you’re interested in what I do there, at some point I’ll link to my PhD student page (when I make it, ha!).

Personally though, I’ve noticed that I’ve started to use data collection to attempt to improve on my life in certain ways.

 

As somebody who enjoys the benefits of e-commerce, I order a lot of stuff online.

This usually means lots of packages. That need to be delivered.

USPS is usually pretty good about being here at a certain time, probably due to having to stop at every single address in the route. UPS and Fedex tend to have much more variability in their times. That variability is a problem when signatures are required. This is especially common if I order something EMS from Japan. And sometimes, you really, really want that package today.

 

Delivery is hard.

Delivery is hard.

So what I started doing almost a year ago was record the carrier, date, and time for each of my packages. In case you’re wondering about n (I’m admittedly not as good about recording the data as I should be):

USPS: 18

UPS: 11

Fedex: 11

I’m missing DHL and Prestige (which is bad since I order a bunch of stuff using amazon prime shipping nowadays).

USPS is relatively consistent, with an average delivery time of 2:11pm and a std dev of 50min

UPS (as expected) is much more variable, with an average delivery time of 1:38pm and a standard deviation of 3hrs and 17min.

Fedex (also as planned) is similarly variable, with an average delivery time of 2:24pm and a standard deviation of 2hrs and 59min.

 

That’s right, a roughly 3 hr window for UPS and Fedex.

 

Now, how does this improve my life?

 

Uh…

 

Well, if I really need to be home for a package from UPS or Fedex, I more or less need to stay home most of the day. That’s… about it. I didn’t say “improve my life in useful ways”.

 

For more data porn, I recently finally read Freakonomics (thanks to my used book buying binge at Powell’s City of Books in Portland). Also, I’ve recently discovered OkTrends which is amazing.

 

Astute readers might remember me trying to collect data on my own habits. I have about 2 weeks worth of it sitting there, waiting to be analyzed. Whenever I do so, you guys will be the first to know.

Also, I’ve started trying to create a data set out of 20+ replays of me playing (mostly losing) Rachel online in Blazblue CS2. I want to eventually cut through that and look at things that happen often, like what I get punished for often and what I tend to hit with. Also, what is the cost (of being punished) and reward (of getting a hit) in terms of damage. If I get anything interesting (maybe even if I don’t ;), I might post about it.

Posted in Life.

Tagged with , , .


3 Responses

Stay in touch with the conversation, subscribe to the RSS feed for comments on this post.

  1. Chuan says

    Uiharu? =________________=

    Great post, though.

    However, after my last research group, I feel compelled to be snobby and point out the use of standard deviation as somewhat misleading. There’s some scientific communities that really stress the reporting of 2 standard deviations 0_o

  2. The Bonkler says

    2 standard deviations? How does that work? Unless you mean mean +- 2 * sd?

  3. still no Uiharu figure says

    Uiharu is the best.



Some HTML is OK

or, reply to this post via trackback.