Data analysis on movie based on Naver Movie to find whether the movie related comments influence the South Korean market compared to other global movie markets.
Movie industry is huge. However, movies make a lot of profit from one country, while it does not in other countries. There are many reasons to this such as culture, timing, population, etc. Out of these, I wanted to experiment to find out if the movie review comments has any effects on the movie's profit in a specific country. As I wish to only figure out a movie review platform's impact, I neglected other components that could affect my experiment such as population by dividing the total profit of a movie by its population. I also compared South Korea with G20 and OECD countries as I target on ignoring the data from developing countries.
The Naive Bayes classifier accuracy was 68.59%, meaning that Naver Movie review platform did not impact a movie's profit. This could mean other countries are same as well. However, there are many uncertainties in this experiment, meaning that other experiments could go more deep into it and improve on this result.
pip install -r requirements.txt
python scrape_movie_stats.py
- Movie Name
- South Korea's Ratio
- Other Countries' Ratio
- Result (1 or 0)
- x = South Korea's profit
- y = South Korea's population
- x = Other countries' profit
- y = Other countries' population
- n = Total number of other countries
python preprocess_movie_stats.py
python scrape_movie_comments.py
python train_test.py