readyresearchers / movies123forme

a streamlight based tool that will provide personalized movie recommendations to users through cosine similarity and text vectorization

Abstract

By looking at the emerging trends in the entertainment industry, it can be inferred that individuals are becoming increasingly interested in the deep personalization of their movie choices. Most streaming platforms now utilize different comprehensive algorithms that keep track of user’s choices, which should further perpetuate the need for unique streaming. However, the research done in thus far has only touched on the different factors affecting movie success and not the rising demand for individualization in streaming. This knowledge gap requires one to explore different ways for streaming services and other sectors of the movie industry to tailor their services for their consumers. This is where the research done in this paper comes in, where the analysis of streaming platforms (Netflix, Hulu, etc.), movie production (actors, budget, directors, etc.), and overall revenue is used to determine what exactly makes a movie successful. This paper will be a comprehensive analysis of different determinants that affect movie success, such as actors, genres, production budget, movie sales, and directors. These determinants will be analyzed with a Logistic Regression algorithm, which will then be put into a machine-learning model. Once the model is appropriately trained, then two outputs will be generated by the model. For movide industry personnel, there will be a streamlit-based application that will allow them to fill out the factors that are being determined by the model with their personal data and then output the liklihood of their movie being successful, based on movie trends determined by the model. For regular users who would like personalized movie recommendations, there will be an interface in the streamlit-run application that will allow users to input their preferred data for the movie factors, with the output being a list of all of the different movie recommendations that the model chooses, based on user preferences and movie trends.

Installation

To run through Streamlit Cloud, simply navigate to the following URL. If you prefer to host locally to your machine, please go through the following installation steps.

As this project relies on Poetry for dependency and package management, the way to install all necessary dependencies is by simply running the following:

poetry install

You should get an output similar to this once all of the packages have been downloaded to your virtual environment:

Installing dependencies from lock file

Package operations: 0 installs, 6 updates, 0 removals

  • Updating attrs (22.1.0 -> 22.2.0)
  • Updating numpy (1.23.5 -> 1.24.2)
  • Updating packaging (22.0 -> 23.0)
  • Updating setuptools (65.6.3 -> 66.0.0)
  • Updating charset-normalizer (2.0.4 -> 3.0.1)
  • Updating cachetools (4.2.2 -> 5.2.1)

Installing the current project: Movies123ForMe (0.1.0)

From there, you can run the streamlit application using the command below:

streamlit run main.py

This should direct you to the homepage of the application.

In the homepage, there are 2 different types of predictive dashboards that you can use:

• Movie Recommender
• 

If you would like to add your own movie to our database, please perform the following steps:

1. run the streamlit command locally using streamlit run mani.py
2. navigate to the 'Movie Search' subpage
3. type in the name of your desired movie. If you see the summary details of the movie, then it has been successfully added to our database!

To be able to get recommendations for your newly added movie, please perform the following steps after you have added it to the database:

• exit the current streamlit session using CTRL + C
• re-run the streamlit command streamlit main.py
• navigate to the 'Movie Recommendation' page and begin getting recommendations for your newly added movie!

---

• Predict Success of Your Own Movie
• 

---

There are multiple subpages included in the application, as well as different features that can be explored on each subpage. These are the different subpages offered:

• Analysis subpage:
• 

---

• Analysis upload subpage:
• 

---

• Regressions subpage:
• 

---

• Machine Texting subpage:
• 

---

• Machine Learning subpage:
• 

movies123forme's Issues

feedback

Some ideas to consider as you begin your next deliverable.

• Polish your final document to resize or remove some of the floating images, anchoring these images to the text.
• Keep a consistent research notebook with current ideas.
• Maintain written tool documentation in the source repository.
• Maintain regular commits for your work.
• Maintain and update versions with your commits.
• Be sure to provide feedback for the work of colleagues.

feedback

In today's CMPSC600 meeting, we discussed the due date for chapters 1 and 2: 20 December at 11:59pm. You will want to add a new release of your document, which has an updated with version number. Each modification of your document will have a new version. Please note that all the baseline requirements may be found in the README.md of your repository.

Some quick (and likely messy) feedback:

• Read over the README.md file to see how to install all necessary software to be able to compile a working release of your work.
• You will have to give the document a version number and a PDF release (see the Section Tagging of the README.md) to learn more about how to do this.
• Build an outline of your ideas to determine a logical flow of ideas in your manuscript.
• Add supporting works from the literature to your ideas of your work. For this, find articles containing discussion of algorithms, theory technology and other parts that will help you to explain and support your own ideas.
  *These cited examples from the literature may be incorporated into discussion from the Introduction and then then expanded upon later in your Related Works Section.
• You will note that there is a .bib file in your root directory into which you are to add BibTex code for each reference.
• The BibTex code comes from many online literature search engines including https://scholar.google.com/. At Google Scholar, for example, find the "Cite" citation button below a search result to find the BibTex code that resembles the following.

@article{lee2022mitochondrial,
  title={Mitochondrial PARP1 regulates NAD+-dependent poly ADP-ribosylation of mitochondrial nucleoids},
  author={Lee, Jong-Hyuk and Hussain, Mansoor and Kim, Edward W and Cheng, Shang-Jung and Leung, Anthony KL and Fakouri, Nima Borhan and Croteau, Deborah L and Bohr, Vilhelm A},
  journal={Experimental \& Molecular Medicine},
  pages={1--13},
  year={2022},
  publisher={Nature Publishing Group}
}

• In your markdown thesis file, you will add the reference key in the following way; [@mycitationReference]. This will to connect the body of text to the reference and cite the article as a number which will correspond to a numbered of references in an automatically generated Bibliography Section. Note: In your text, you would include [@lee2022mitochondrial] to connect the reference to the text.

• Please let me know if you have any trouble with making this code work.