Comparing Web Archives by Using Large Numbers of Images

In my last post, I walked people through my thoughts as I explored a large number of images from the Wide Web Scrape (using, as noted there, methods from Lev Manovich). In this post, I want to put up three images and think about how this method might help us as historians. Followers of my research might know that I am also playing around with the GeoCities web archive. GeoCities was arranged into neighbourhoods, from the child-focused EnchantedForest to the Heartland of family and faith or the car enthusiasts of MotorCity. Each neighbourhood was, in some ways, remarkably homogenous.

Let’s take every JPG from the ‘Athens’ (the teaching/philosophy/etc. focused area of GeoCities) and see what we find.

The Athens neighbourhood, around 5k images. Click through for a relatively low-rez image, and let me know if you have a research need for a high-rez one.

The Athens neighbourhood, around 5k images. Click through for a relatively low-rez image, and let me know if you have a research need for a high-rez one.

Compared to my last post, we certainly see less pictures of people, although they’re probably still the majority of posts. There’s a lot more clip art, low-resolution images, etc. We see some foreign writing, but more importantly, a lot of other pictures: travel adventres, space-related photos, architecture-related photographs, graphical elements used for site navigation, history, etc. Even looking at it from afar, we can see a bit less colour variation: a lot of white backgrounds, for example.

Let’s look at another neighbourhood, this time the child-focused EnchantedForest:

The EnchantedForest neighbourhood, around 5k images. Click through for a relatively low-rez image, and let me know if you have a research need for a high-rez one.

The EnchantedForest neighbourhood, around 5k images. Click through for a relatively low-rez image, and let me know if you have a research need for a high-rez one.

More colour, although still a lot of black and white images: closer inspection reveals that it is an electronic components focused website, taking up the roughly second quintile band of the image. Otherwise, lots of children’s characters, site navigation elements, cartoons, colour, pictures of pets, babies, dogs, etc.

In conjunction with textual analysis, I think we can learn a lot about a neighbourhood – at first glance – by using these photographs. In any case, it’s an interesting visualization to explore.

Finally, if we look at the montage from last week, we can see how things have changed between the heyday of GeoCities (many of these sites date from the late 1990s). Digital photography, at a glance, has certainly played an even bigger role of late!

From my last post, the 50k montage.

From my last post, the 50k montage.

One thought on “Comparing Web Archives by Using Large Numbers of Images

Leave a Reply

Please log in using one of these methods to post your comment:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s