Conversational Analytics on Hotel Data-Turning 50K Reviews into Insights with Wren AI

No complex schemas. No SQL. Just your data - and instant answers! From Code Blocks to Conversations

I recently came across a rich synthetic dataset from Kaggle that simulates the booking and review data of a global hotel platform.

https://www.kaggle.com/datasets/alperenmyung/international-hotel-booking-analytics?resource=download

When I first saw this Kaggle notebook, I instantly recognized the familiar rhythm of a data analyst's workflow: Setting up the environment, Importing libraries, and preparing to explore a dataset through code. 

Instead of firing up Jupyter and writing code, I turned to Wren AI - an conversational BI platform that lets you do the same analysis with plain language. 

It's a perfect sandbox for learning - a dataset that's synthetic yet structured enough to simulate real business intelligence and customer satisfaction modeling.

🗣️ The Wren AI Approach - Ask, Don't Code

I wanted to try something faster by uploading the CSVs directly into Wren AI and exploring insights instantly through conversational analytics.

1. Go to getwren.ai.
2. Upload the files: hotels.csv, users.csv, and reviews.csv.
3. Wren automatically detects relationships between tables (hotel_id, user_id)

Within seconds, you'll see your dataset appear as an integrated workspace ready for analysis, instead of writing pd.merge() or groupby, I can just ask questions like:

"Top 10 cities have the highest review scores?"

Ask follow-up questions like,

"Compare review scores by traveler type."

The chart showed that solo & business travelers consistently give lower scores, while couples and families tend to rate their stays higher.

"Break that down by traveler type"

or

"Show top 10 hotels in Asia."

🔍 Why Wren AI Changes the Game

Traditional BI requires setting up schemas, joins, and dashboards. Wren AI translates each question into optimized SQL, executes it, and visualizes the results - all in one conversational flow. What makes this dataset exciting is how relational it is - the three tables connect naturally, allowing for deeper, multi-angle exploration.

"Do 5-star hotels outperform 4-star ones?"

"Which traveler type gives the lowest ratings?"

"Which hotels are underperforming based on their star rating?"

📊 Insights I Discovered

Correlations among review metrics (cleanliness, comfort, staff, location) revealed that staff quality and cleanliness were the strongest predictors of a high overall score - stronger than location or even price.

"Generate reusable Insight or visualizations directly for presentation or reporting."

"Show me average review score and number of reviews per hotel."

"How does traveler age impact hotel choice (based on star rating)Score_staff vs. Score_cleanliness."

It's like having a data analyst that already understands your dataset - and speaks fluent SQL behind the scenes.

⚙️ Under the Hood - Wren's Semantic Layer

Behind every question, Wren AI automatically:

1. Maps entities like hotel, user, and review into structured relationships.
2. Understands metrics (scores) vs. dimensions (city, traveler type).
3. Generates the necessary SQL logic on demand.

It's essentially the same intelligence a data analyst applies manually - but expressed through natural language.

Wren AI doesn't replace Python notebooks - it democratizes them. Data analysts still benefit from full control when needed, but business users, marketers, or product leads can now explore the same dataset without technical barriers. 

It bridges the worlds of analytics coding and AI-assisted BI - turning complex data into approachable insights.

🌟 Final Thoughts

From Raw Data to Ready Insights - In Minutes Whether you're in hospitality marketing, hotel management, or guest experience for international brands, Wren AI transforms CSV uploads into business intelligence - supporting decision-making. Ready to turn hotel reviews into competitive insights?

👉 Start a free demo at getwren.ai and see actionable data in minutes.