Mini Project Ideas

[oliviapetronio]

No description provided.

[oliviapetronio]

Data 400-mini project.pptx
Data 400-mini project.pdf

[ernbilen]

• The genre comparison idea is great—explore differences between metal, country, and pop by gender. Could compare earnings or streaming algorithm patterns.
• Consider looking at trends over time (e.g., 80s, 90s).
• Creative visualizations are encouraged—check out r/dataisbeautiful.

[oliviapetronio]

Progress Presentation.pptx

[ernbilen]

• Use a better (more visual) title slide :)
• Add two more predictors to your model, that are, distance to the nearest target warehouse, local population within say 20 mile radius. (I like your distribution center slide)
• Another predictor you can use is gas price.
• Use some python API to get CPI, that should be easier that BLS finder platform.
• Your milk price slide was the most catchy and attention catching side for sure ✅
• Egg price was interesting too.
• Of course in the future instead of screenshots you can use graphs to show prices. like histogram etc. also you should probably get data from different regions too, south, midwest, west etc.
• As a sanity check, you can try showing electronics too, and that prices are same. So maybe the story is about perishable food? Are you seeing price differences in any other categories? How about non-perishable food like canned food? Are they same across places?
• Your urban, suburban, city col is great. Consider adding rural locations too.
• For your model, it makes sense to consider inference using linear regression. Are prices cheaper in rural areas etc. Clustering is great too.
• Great use of maps overall.

[oliviapetronio]

ReadME

[oliviapetronio]

Data 400 - Final Presentation.pptx

[ernbilen]

• Nice intro with questions for audience.
• I'd consider changing your font style, pick something more futuristic or "data-sciencey."
• I'd use a different point style for distribution center. Maybe a larger point compared to stores?
• Your project is set to know why say central pa has more expensive prices at the store right? That is, checking correlation between price and say demographics. If your model performs poorly they you can say "we don't know" based on the model.
• Before breaking results down by region or income, you can start by showing the raw $ cost of the basket and show how price differs. Which I think what your map shows, so maybe you can reverse the order and show the map first.
• You can include all items in the same regression in your models with cost of living as the y variable.
• You tied your implications well with the results of your project, you should do the same if possible for the ethics discussion.