Boardgame-Market-Demand-Forecast-with-NLP

Dataset can be retrieved from https://1drv.ms/u/c/b574be9cdd5eafe3/EW_LpyLiAPtApitiXRo_a3YB4KzDj9EwRlN66vdCKTNjDQ

Project Goals

• Develop a model to analyze the characteristics of popular board games.
• Identify key attributes and mechanics that drive sales.
• Use insights to evaluate and improve GameTrend’s existing game portfolio.

Feature Engineering

• Regular Expressions: Used for cleaning and structuring raw comments.
• Using models like GPT-3.5 Turbo for summarization and difficulty detection.
• Leveraging models like BERT and XLM-RoBERTa for sentiment analysis.

Modeling

Used various regression models for sales prediction of the board games.