search

Iveynganga / Movie-Recommender-System-Capstone-Project

0 stars 0 forks source link

Review 2 #2

Open okothchristopher opened 1 day ago

okothchristopher commented 1 day ago
  1. On the Readme, this section here should have small headings : image And possibly be indented. The same comment goes to the bit on Cosine similarity.
  2. Problem Statement - Will this recommender system be able to amalgamate movies from other platforms , e.g Should we expect to see a recommendation to a movie in HBO if I only select Netflix Based Movies? Because I think this is the main problem that people would want to solve with a system. Otherwise someone will say, I already get good recommendations on Netflix anyway, why should I use a website for this, if it is not going to tell me about similar movie on other platforms.
  3. Good work on the Readme. It makes someone understand the project end to end.
  4. What you should also include is a link to the Notebooks and Files : Use this template for your Readme formatting.
okothchristopher commented 1 day ago
  1. I cannot see a file for the data
  2. I cannot see a file for the presentation
  3. Also include what else are you doing. What is going to be the end output of this project. You still need to document what you will be achieving in the future. Which should be before the 12th of October. Deployment with Streamlit will be very critical for your project since it is something that includes a user.
okothchristopher commented 1 day ago
  1. You also need to include interesting statistics in the Readme, from your EDA.
  2. You did not tell us about the data source.
okothchristopher commented 1 day ago
  1. You need to clean the data better.
  2. Here you need to split these and use a count vectorizer. image
  3. There are no interesting statistics about your data. For every graph you do, there has to be some explanations in form of insights, or deductions. We talked about these, you need to even tell us statistics about the movie release dates. Which movies have the highest gross margins, profitability etc. in essence you need to perform more EDA.
okothchristopher commented 1 day ago
  1. You need to surpress all the warnings.

  2. image Once you are done with the explorations of your data, leaving these outputs do not help the reader. image

  3. Why am I seeing EDA in the Modelling Notebook.
  4. Why is there no notebook for data sourcing.
  5. You need to explain the final result, why do you think movies similar to Ice Age are the ones listed. Why not test with other movies as well.