Create Mini Project

[Nishthas05]

…- Nick Singh, Kevin Huo -- Place of publication not identified, 2021 -- Ace the Data -- 9780578973838 -- 8ffbabf64bf797ac2765e98282099a59 -- Anna’s Archive.pdf

[Nishthas05]

Data Senior Seminar Mini Project Proposal.docx

[DerekChibbaro]

Really cool idea and I like your thought process behind the variety of models that could be used to analyze how the naturalization rates are correlated with historical events. If major events seem to be outliers, you could group them all together to see how the major events vary (maybe try to see which major event causes the highest or lowest spike in naturalizations?)

[ernbilen]

I like this a lot. EDA-wise you have a lot to explore, trends, patterns.. I don't think your logistic regression makes sense in this context, but k-means does. I'm not familiar with how naturalizations work, as in, if there are country based caps, but one thing you can explore is h1b or green card caps based on a country and how those caps are allocated. Another thing to keep in mind is naturalization is a choice, as in, after 5 years of being a green card holder, you have the option to apply for naturalization, so keep in mind when interpreting your results. Also, I wonder if you have access to denial rates. Lastly, you may need to go beyond Kaggle's data and get data from USCIS directly.

[Nishthas05]

Data Senior Seminar Mini Project Proposal 2.docx

[DerekChibbaro]

I think something to look at is the time of year certain events are more heavily covered versus others. Are there certain types of events that are talked about more in the summer or winter and why would that be the case? You could also think about how different events are talked about using your text analysis...negative or positive? Many news outlets are usually biased in some ways so you could analyze how they compare to each other.

[Nishthas05]

https://docs.google.com/presentation/d/1mFM7puio23kCKcb4SRdJtKQ9mv1V8Ul_Tf7Z3FMj_Fw/edit?usp=sharing

[ernbilen]

• Memes are always appreciated.
• Nice sample range!
• Consider using USCIS data directly.
• For exploratory data analysis (EDA), try k-means clustering at the country level.
• The region-wise chart is useful.
• USCIS categorizes some countries based on backlogs—could this be leveraged?