Conversation
|
Most updated version: |
|
For Mini Project 1 I really enjoyed reading your plan for this project. I think it is very detailed and it should be interesting to see what results you get from your exploratory analysis. This is a very strong project that incorporates a lot of strong economic and business terms. Be sure, when presenting this project, that you define and describe the purpose of these terms before diving into the project just in case your audience does not have an economic or business background. I like your plan for EDA and your model, but I would like to see, if you have the time, a stronger predictive model than regression. Maybe something along the lines of gradient boosting or random forests. Also, instead of just sticking with some macroeconomic subjects, maybe expand your indicators to other classes. Maybe look at other indicators that huge finance firms consider when making decisions or analyzing the market. My last suggestion is that if you do decide to do a predictive modeling, I would keep an eye out for huge outliers such as the COVID-19 pandemic and the 2008 recession. These could potentially be very helpful in your predictive model, but they also have the situation where they can skew your prediction. Otherwise, you have a really strong project and I really look forward to seeing the final product! |
|
Like your Mini Project 1, your second one is very put together and has a nice end goal. It is very detailed and I am excited to see the end result. Similarly to project 1, be sure to explain some macroeconomic terms to your audience as it may not be obvious to them on what certain business terms are and their purpose. To enhance your project, I would suggest to also look into inflation rates and unemployment rates during the time span you are looking. These should also be available on FRED and it will also give you more insight if the yield curve is a strong indicator that a recession is coming. A regression model is a strong choice for predicitve modeling, but yield rates are not necessarily linear. I would check if your data is linear, if not, I suggest looking into predictive models that are well-suited for nonlinear data. If you can make a logistic regression model work for the nonlinear data, that is great, but I would just be wary and double check that the prediction is well-fitted. I also look forward to seeing the results of this project, if this is the one you choose to do. |
|
Nice line charts, but I think monthly might look a little better? |
3 participants
Mini Project 1 - Thanh Vu.md