OUR DATA

For this Project we used the transcripts from the DanDaDan manga and anime as a base, going in with XML (specifically with TEI and CBML) to label each line of text. We found 3 things important to label here:

• Names: Very obviously important, we labelled each character with consistent character names. In the manga transcripts we encoded with attribute who on speech bubble elements. In the anime transcripts we encoded with both speaker elements and who attributes.
• Diegetic Text: There are magazines and location signs in this media, and we wanted to compare the two. In the manga transcripts we encoded with element cbml:caption. In the anime transcripts we encoded with caption elements.
• Scenes: To help divide the manga and anime into easily digestible segments we chose important points to "break" the text into, making it easier to analyze in smaller chunks. In the manga and anime transcripts we encoded with div and id elements.

Transcripts of the anime already existed however for the manga we had to manually transcribe all of the text which proved to be a time-consuming task. When you read manga you read it right to left, which is a small hurdle when reading as a human, however this meant that we couldn't directly copy and paste the text, as due to the reversed order and wonky panel structure, it would paste totally out of order. As a result we had to manually go through and paste each and every bubble individually!

Let's look at an example of a particularly text-heavy page (click the manga to hear audio from the anime!):

If you noticed and are wondering why we didn't bother blocking out the panels in the encoded manga, the answer is quite simple: our goal is to compare the manga and the anime, not just to encode the manga to a tee. The anime doesn't have panels in the same way the manga doesn't have sound effects (obviously there are comparably alternatives in both, but none that would easily work for the current scope of our project). Check out the scrollable sections below if you want to see the full medias that we worked with, or follow this link to our GitHub!

For further information about what we found in this data, check out our Analysis Page.