AI-Powered Predictive Analytics: Driving Smarter Decisions for Peek Travel
Experiences and AttractionsBusiness Impacts
Attained 92% accuracy in sentiment analysis using the advanced Gemini 1.5 Pro model, driving precise customer insights.
Boosted demand forecast accuracy to an average of 70%, with peak performance reaching 82%, enhancing decision-making.
Developed Looker Studio dashboards featuring 20 actionable KPIs, enabling effective visualization of key model insights for data-driven strategies.
Customer Key Facts
- Country : United States
- Industry : Experiences and Attractions
- Website : www.peekpro.com
Problem Context
Peek Pro provides world-class online booking, point-of-sale, and hundreds of automation tools such as inventory management, dynamic pricing, waivers, and marketing analytics. Thousands of operators have supercharged their revenues and automated operations with this all-inclusive technology.
Peek Travel sought to enhance its solution portfolio by introducing a suite of predictive analytics tools designed to provide its partners with deeper operational insights. However, Peek’s existing demand forecasting model did not effectively account for seasonal trends and variations in customer behavior and transactions. Additionally, they faced challenges in extracting sentiment and preferences from customer reviews and lacked a robust system for delivering actionable insights to their partners.
Challenges
- There was a need to enhance the existing demand forecasting models to capture seasonal trends and shifts in consumer behavior.
- The vast and complex nature of customer review data made it challenging to fully leverage it for enhancing service offerings.
- A more streamlined system was needed to deliver clearer, more actionable insights for business stakeholders.
- Refining forecasting capabilities presented an opportunity to enhance the reliability of predictions and strengthen data-driven decision-making.
Technologies Used
BigQuery
Vertex AI
Looker Studio
Cloud IAM
Cloud Source Repositories
Gemini
Solution
Quantiphi implemented a comprehensive predictive analytics solution with three key components:
- Demand Forecasting and Price Optimization: A machine learning-driven demand forecasting model was developed using time series analysis to predict future demand for specific partners. This model was integrated with a price optimization system that leverages demand forecasts, historical booking data, and business logic, enabling more informed and strategic pricing decisions.
- Sentiment Analysis: Quantiphi developed a sentiment analysis tool that leverages the Gemini 1.5 Pro model to categorize customer reviews and extract valuable insights into customer sentiment and preferences, enabling better service enhancements.
- Data Visualization with Looker Studio: Interactive dashboards were created using Looker Studio to visualize insights, KPIs, and metrics from the demand forecasting, price optimization, and sentiment analysis models, enabling data-driven decision-making and provide actionable insights for the business.
Results
- The predictive analytics solution delivered a substantial improvement over the existing system, achieving 92% accuracy in the sentiment analysis model and approximately 70% accuracy in the demand forecasting model, with performance peaking at 82% for select partners. These enhanced accuracy levels have directly contributed to more precise decision-making, enabling better alignment with customer sentiment and more effective demand planning.
- A user-friendly dashboard, incorporating 20 key performance indicators (KPIs), empowered stakeholders with actionable and interpretable insights and established a strong data foundation for future growth.
