Weekly Project Update 3/30

[ilk18]

Meeting at 3:30 pm in 305 Hillman Ben, Madelyn, Claire and India in attendance, Dr. Birnbaum observing

To begin the meeting, we had a discussion about making edits to our XML schema.

• For our action element, we changed the “who” attribute from "self, opposite, third party" to "man, woman."
• We changed our “type” attribute from "direct, indirect" to "active, passive."
• We also also added "cheating" to our "fault" attribute, as it is common in many of our songs.

We feel very good about the new schema—it is more concise and better fits our research question.

With this updated schema, we are keeping in mind that we hope to include graphs looking at a correlation between description and gender. (Are men, women spoken about more positively or negatively in these songs?)

Then, Madelyn showed us her first draft of the website. Preliminary edits included fixing the song on the homepage so it does not autoplay, and moving the tabs at the top to the left side of the page. India offered to write the biographical and historical context blurbs for the website. We also agreed on our reading view.

For next time:

• Madelyn will continue adding to the website
• India will make a XSLT stylesheet for the corpus
• Claire will start working on Javascript

[JackGutch]

I like all of the insight provided. When looking at the attribute element, our group choose to stick only with "P" for positive and "N" for negative. I like the changes, as I feel we could have shifted the attribute element to "increase" or "decrease" for certain elements. This would give a more understanding and unique approach to certain elements, as positive/negative is seemingly vague. We also hope to include graphs that depict relationships between various elements and their respective attributes, in order to find themes in our corpus of documents. Overall, progress looks great and we are following a similar path!

[mkd38]

I think looking at how positive and negative sentiment breaks down across gender is a great idea and might be very revealing! Especially keeping in mind that it seems that most (all?) of your singers are female. Whatever the results of your analysis, as a suggestion for further research, you might in the future decide to repeat that analysis with a balanced corpus and see if the results change or hold constant. If the latter, that might indicate that these are conventions of the genre rather than reflective of the singers' individual attitudes.