• Trained on a storytelling-style corpus of 10 narrative sentences.
• Uses trigram modeling to predict the next word based on the last two.
• Tokenized with tidytext and filtered using dplyr.
• Prediction logic:
  • Extract last two words from input.
  • Match with trigram dataset.
  • Return most frequent third word.

• Hosted on shinyapps.io
• Input: Any phrase with at least two words
• Output: Predicted next word from the corpus
• Built with:
  • shiny for UI
  • tidytext for NLP
  • dplyr and tidyr for data wrangling

• “Once upon” → a
• “She would” → walk
• “To the” → river
• “The magical” → box
• “Little girl” → who

Each prediction reflects the narrative tone and structure of the corpus.

• Enhances writing tools, educational apps, and creative assistants.
• Lightweight, fast, and easy to scale.
• Future potential:
  • Larger corpora (news, social media, books)
  • Multilingual support
  • Deep learning integration
• Let’s build smarter storytelling together.