1. Which dataset did you use?

Multiple datasets collected from publicly available sources, including Yahoo Finance, Google Maps, and various corporate financial documents

2. How did you analyze or prepare the data?

• For the hotel review data:
  
  We built an end-to-end pipeline that collects hotel review data from Google Maps using a Python script data scraper. Each review record was standardized into an ML-ready format with normalized text, location/property attributes, derived time attributes, and review IDs. Due to OAC data flows and ML algorithms being disabled, we created and ran a rules-based ABSA to simulate OCI-style sentiment analysis that produces both review-level sentiment and aspect-based sentiment with span metadata (offset/length), confidence scoring, and full raw text for traceability and auditability.
  
• For the financial data:
  
  We used a combination of Yahoo Finance to collect the stock data and AI-assisted parsing of corporate financial statements/documents

3. Who is the intended audience for your visualization?

This type of historical analysis - and a similar type with real time or near real time data - would be of interest to the following:

• Brand/Marketing Leaders (CMO, brand directors)
• Operations & Quality Leaders (COO, regional ops, QA)
• Guest Experience / CX Teams
• Revenue Management & Commercial Strategy
• Portfolio/Asset Managers & Owners
• Corporate Strategy / BI / Data Science
• Travel Industry Analysts/Consultants

4. What is your visualization about, and what question or problem does it address?

The objective was to analyze data across the hotel industry to identify potential impacts or trends after COVID. We looked at hotel chain revenues, room counts, and stock data from a financial perspective, as well as conducted a sentiment analysis from user review data from Google.

5. Did you use any Oracle Analytics AI features when building your visualization (ex. AI Assistant)? If so, please describe how they were used.

While we weren't able to use ML features within OAC, we tried to simulate this process using an aspect-based sentiment analysis and then visualizing that data in a tag cloud and text highlight plugin.