Next Word Predictor App Pitch

June 2025

Slide 1: Introduction

Next Word Predictor: A Shiny app for real-time word prediction
Goal: Enhance typing efficiency in keyboards, chatbots, and apps
Data: “The Adventures of Sherlock Holmes” from Project Gutenberg
Try It: https://etr80a-afnan-khan.shinyapps.io/next-word-predictor-app/

Slide 2: The Algorithm

N-gram Model: 3-grams with 2-gram back-off
Data Processing:
- Used full text (~0.6MB) from Sherlock Holmes
- Cleaned: lowercase, removed punctuation/numbers, Gutenberg header/footer
- Robust handling of empty lines and corpus issues
Why It Works: Fast, lightweight, suited for narrative text

Slide 3: App Functionality

Input: Phrase in a text box
Output: Single word prediction after clicking “Predict”
Features:
- Simple UI, clear instructions
- Deployed on shinyapps.io
- Fast response using precomputed N-gram tables
Link: https://etr80a-afnan-khan.shinyapps.io/next-word-predictor-app/

Slide 4: User Experience

Ease of Use: Type phrase, click, see prediction
Testing: Predicted next word for 5 phrases:
- “it was a” → “very”
- “he said” → “that”
- “the door” → “was”
- “i have” → “a”
- “in the” → “room”
Feedback: Intuitive, reliable for narrative-style phrases

Slide 5: Why Hire Me?

Novelty: Compact N-gram model optimized for small datasets
Skills: R, Shiny, NLP, data preprocessing
Impact: Ready for integration into real-world applications
Hire Me: I bring technical expertise and innovation to your data science startup!