search

nickeubank / mtv_viacom_capstone

1 stars 0 forks source link

Final report plan #47

Closed nickeubank closed 2 years ago

nickeubank commented 2 years ago

Greg says I'm supposed to grade a final report plan? Can you pass it along please?

jgy4 commented 2 years ago

Ah yes! We uploaded our Major Semester Milestone plan here:

https://github.com/nickeubank/mtv_viacom_capstone/blob/main/40_docs/Major%20Semester%20Milestone.pdf

nickeubank commented 2 years ago

Perfect, thanks!

nickeubank commented 2 years ago

Thanks for this!

Reading this over, I think you've done a very good job of meeting the requirements of the assignment (articulated some very good concrete objectives, all very relevant to the client's needs). However it is my sense that some of these goals might not be the best prioritization of your effort.

I think the main objective that we should have over the next couple weeks is really making sure that we understand the data that were working with and all of its potential issues that so that we can provide answers to a couple of the clients questions with really high certainty -- namely Question 1 and Question 3 in your report. Also well I think it is going to be important to have some good visualizations, whether we need to convert those into a dashboard at this point feels like an open question to me -- I will admit that in general I feel like dashboards are kind of a default deliverable that people are into, but in general I find their value dubious.

Based on our conversations as a team in the conversation Adriane I recently had with Vaughan, it sounds like they're real clear priority is to have a couple topline empirical findings that they can include in a press release, and while that may not feel sexy as compared to a dashboard or somebody showing off the model they claim performs well, I think the reality is that having a couple topline empirical results that we feel sufficiently confident in that were willing to have included in a press release is actually gonna require a lot more work.

Part of the problem here, I think, is that the type of work that we still need to do is probably the kind you haven't had a lot of opportunities to practice in classes -- in classes, we usually have the luxury of nice clean data. But with working with the type of data that were working with, we really have to be absolutely sure that we know every merge and every subsampling backwards and forwards. This is something that Adriane and I will continue to work with you on, but hopefully are starting to get a sense of the types of concerns that I think we should be focused on from my recent code reviews and some of our discussions about data provenance. Were getting there, don't get me wrong, but before we say anything on politically sensitive topics and press releases we really need to make sure that we understand our data, and were not there yet, and getting there will take enough work to keep us busy to the end of the semester.

With that in mind, I would suggest the following as end of semester goals:

I recognize that some of these are probably less specific than Greg and Ryan would like, but one of the challenges of this kind of analysis is that resolving unknowns is both the hardest and most important thing, and it's hard to fully specify how we'll do so in advance. I think @adrianefresh and I, for experience, have a good sense of what remains to be done in this area, and we will keep pushing you with specific questions where needed (though hopefully you're starting to foresee these!), but until those are all done, I'm not sure investing in something like a dashboard or a Shelby v. Holder analysis is appropriate.

jgy4 commented 2 years ago

Thank you for this feedback Nick! I think this all makes a lot of sense as we narrow in on the rest of the semester.