Weekly Project Update 2/23

[mam797]

Weekly Project Update 2/23

India, Claire, Madelyn, and Ben were all present at the meeting.

Discussed

• Midterm
• What comes after the midterm
• Our work on our schema
  • Add <lb/> so you can have an empty element span lines
  • Add <br/> for every time you encounter new element in xhtml, start each line where you have <lb/> element
  • Line-start needs to be pulled out of a string of elements because there could be multiple line starts
  • Fix attributes so they are consistent
  • Combine third-party person/third-party object to become a third-party entity
  • Change man to the opposite party in terms of people involved?
  • Maybe add family issues under the cause?
  • Keep x (placeholder) for when we run into unexpected tags
  • Change money value to static if it doesn’t relate to gaining or losing
  • Current quantity of money is related to lyrics directly
  • Change money amount to lacking or enough
  • Make money loss/gain optional
  • Delete one of the schemas (we have two on git)
• Our work on our methodology
  • Need to write in <lb/>
  • Add all of the above so that songs will work with schema
  • Combine India and Madelyn's methodologies

    Assigned for the midterm

• Maddy has 6 Bessie Smith songs and Slave to the Blues
• India has 6 Mamie Smith songs and Stack O’ Lee Blues
• Claire has 3 Ma Rainey songs
• Everything should be in the repo by the midterm date
• Peer review after midterm, but try to do as well as possible

  Things to think about

• Do we want to use topic modeling? We should answer this in slack after reading Mining the Dispatch.

[mkd38]

Interested to know the decision that you take on topic modeling. Personally, I'm conflicted on whether our team ought to use it to enhance our analysis. Topic modeling is less useful in short texts, like songs (or surahs) since short texts can be analyzed through close reading, but at the same time, you could use topic modeling tools in really interesting ways on your corpus (e.g. you could review proportionality of topics across all artists in your dataset or something). My thought is that it might be good for our team to analyze our surahs through close-reading but apply topic modeling to our full texts, so that we can say we compared the texts on small scale and large scale, which might make for a more valid analysis, but we are still in discussion about this.

[JackGutch]

I like all of the progress you have made so far. Within our group as well, much of the weekly meeting conversations have been centered around fixed the schema so that it is perfect before we get too deep into the markup of the text. We have added and taken out many schema elements and their attributes as well, in order to show change over time or change in a period of documents. Our tentative due date for finishing the markup is the same, and thus you should be on a good pace towards finishing in time. We are yet to discuss topic modeling as a group, but I personally think the addition of it in your project would be helpful.

[pamelarsmith]

I find it interesting to see that you changed several attributes to be more value-neutral; person/object to entity, money to static. We had to slightly do this too when figuring out what relationships, actions and emotions to use in our markup, to reduce crossover. I think it would be really interesting if you used topic modeling in your project. I think this would fit pretty well because it sounds like you are already marking up lyrics based on topics and themes.