Visit (Jordan)
- I really like the concept of this project, it seems like a great area to use linguistic data science with. The way you collected the best rappers of all time is great, when you were first telling the class about this project I was wondering how you'd choose which rappers to include, but a community-made list seems like a great choice. It'll be very interesting to see how the themes expressed in rap change over time.
- Since your project is specifically about the black experience, I wonder what your methodology is going to be to filter out rappers who aren't black. In addition, you might find some issues in groups like Bad Meets Evil where it will be difficult to separate Royce da 5'9's lyrics from Eminem's lyrics.
- I learned that Genius has an API, that sounds like a great resource compared to doing web scraping
Thank you for the compliments! I think that it will be very interesting to see what I find in the analyses. I filtered out non-Black artists by personal/attested knowledge of the individual Rapper ethnicities. Genius's API is so interesting to work with.
Visit (Frances)
- Your project is very well annotated and easy to follow through the comments.
- I think you mentioned it, but cleaning the data could be a good idea (although from what you included it doesn’t look too bad, just some \n characters).
- Interesting that Genius has its own package for lyrics!
Thank you. I try to annotate as much as possible because I, myself, get confused. I want to make sure future me can understand what is going on too. I found some feature data in there and I'm not quite sure how to filter that out as well. But maybe I won't have to? Stay tuned.
Visit(Emily)
- This is a really interesting project and I look forward to the presentation! I am impressed how you worked around the roadblocks in collecting your data, I know from experience that can be a longer process than anticipated!
- What tools will you be using to complete your analysis? Are the specific thematic representations and words they use something that is commonly ascribed to rap or will you make you own list? I know this is all to come but I am curious as to how exactly you are going to achieve your goal.
- I learned about the importance of utilizing others code to further your own project.
It took so long to do it but all in the name of research I suppose! I will most likely use word clouds after I do LDA topic modeling since I didn't gather Billboard year/rank data. I will then use the word clouds to make wide generalizations. I may include some cluster data scatterplot graphs as well. I'm not sure just yet. I'm still learning that one does not have to reinvent the wheel.
Visit (Emma)
- Your process markdown file was a really cool idea to show the way you went about collecting that data. I also really liked how thorough you were with your data collection, it was really clear how much you are interested in the analysis of your topic!
- There still seems to be more you have to do to start your analysis, but you note that in your 10th log, so I’m sure you have that planned out!
- pip install lyricsgenius
Thank you! I'm glad you were able to pick up on how passionate I am about this project. I still have a long way to go and am hitting plenty of roadblocks code-wise. I will conquer and prevail!
Visit (Abby)
- This project is a very interesting exercise in sociolinguistics. I absolutely love the title and concept.
- Is lyricsgenuis able to scrape other data? In the project plan, you mentioned comparing lyrical topics by year, and I think that was a great idea (but I very much understand if plans had to be scrapped once the project actually got going). Since male versus female is a data point you already have, it would be interesting to do statistical comparison of male versus female topics.
- As Emily mentioned, building from others' work is extremely helpful. There's a lot of already-written code out there!
Thank you so much! I try to be as engaging as possible with my titles. Though the saying goes: "don't judge a book by its cover"; we do exactly that. So I try to make the cover as interesting as possible. Statistical comparisons are exactly what I will be going for!
Visit (Sonia)
- It's a very interesting project and I like how well-documented you are about it, especially in the progress reports and frequent log entries
- I don't have much critique so really a nitpick: in lyrical_data_analysis.ipynb, lines 3 and 4 of the dataframe at the bottom look to be nearly the same song (maybe a remix)? I wasn't sure if this was a common issue or not in the data but maybe something to look into if you have time
- I didn't know reading from a file could be so difficult, glad you were able to figure it out!
Once I figure out how to populate the darn dataframes themselves, I will try my best to filter out repeat information. Not sure how I will do that yet. There are a lot of things I will need to filter out such as 'Skits' or 'Interludes'.
Visit pt.2 (Frances)
- Your data collection is great and your log is super updated which made it easy to see your progress!
- It sounds like you have your data mostly ready to go for analysis, and so you might be doing this once it’s cleaned, but are you going to include more stats like token count (unless that’s not relevant to your analysis lol).
- It’s interesting that the lyrics for different artists are all stored in different json files.
Thank you, Frances! I will be doing this once it has been adequately cleaned. Hopefully that won't take as long to do so I can finally start doing some basic stats then start doing the LDA analysis and the word clouds.
Visit (Michael)
- Your analysis notebook is incredibly well documented and easy to follow! I also appreciated that you included the process by which you gathered reviews so that others could do the same.
- Not really much to say in the way of critiques. It could perhaps be a little clearer, either in the analysis notebook or in README.md, where the code that was used to scrape the reviews came from.
- From the quick glance I took at the API client, it looked pretty interesting.
Thank you, Michael! I'm glad it was easy to follow. Thank you for the suggestion, I didn't realize that I didn't include this information. I'm hoping to fine tune my
README.mdand the other documents today or tomorrow.