Planning a vacation to Zanzibar and want to make sure you book a hotel with good weather and plenty of nearby restaurants? What if you could analyze 11 years of rain data and calculate walking distances to every restaurant on the island—all with a few simple prompts to an AI assistant?
That's exactly what we're going to do in this tutorial using Atlas and Navi, our AI agent currently in open beta.
Traditional vacation planning relies on reading reviews, checking weather averages on websites, and manually looking at maps to see what's nearby. But you're missing the data-driven insights that reveal actual rainfall patterns over a decade and precise proximity analysis for every hotel on the island. That's why smart travelers ask: can we use GIS and AI to transform vacation planning from guesswork into data-backed decision-making?
With Atlas and Navi, you can analyze historical climate data, calculate restaurant accessibility, and create an interactive map that shows you exactly where to stay based on your priorities. No GIS expertise needed, no complex software, no manual calculations. Just conversational prompts and AI-powered spatial analysis.
Here's how to do it step by step.
What We're Going to Build
Here's a preview of what we'll create—an interactive map where you can explore hotels, see rain data, and find nearby restaurants:
By the end of this tutorial, you'll have an interactive map showing:
- All hotels in Zanzibar
- Number of restaurants within a 15-minute walk of each hotel
- Mean rainfall from 2014-2025 for February (your travel month)
- Nearby tourist points of interest
- A beautiful interface to explore and compare options
We'll use free CHIRPS climate data (Climate Hazards Group InfraRed Precipitation with Station data) and OpenStreetMap data for hotels and restaurants. And we'll let Navi do most of the heavy lifting through natural language prompts.
Ready? Let's get started.
Step 1: Get Access to Atlas and Navi
First, head to app.atlas.co and sign up for a free account if you don't have one already.
Once you're in:
- Create a new project or open an existing one
- Look for the option to opt-in to the Navi public beta—you'll find this in your project settings or it may prompt you directly
- Once enabled, you'll see a blue Navi button in the top left of your interface
Navi is Atlas's AI agent that can understand spatial analysis requests in plain English, execute complex GIS workflows, fetch data from OpenStreetMap, perform calculations, and even style your results beautifully.
Step 2: Download Historical Rain Data
For this analysis, we're interested in February weather (a popular time to visit Zanzibar). We'll download 11 years of February rainfall data from CHIRPS.
What is CHIRPS? It's a high-resolution precipitation dataset that combines satellite imagery with station data, providing reliable rainfall estimates globally since 1981.
Here's how to get the data:
- Go to the CHIRPS data portal: https://www.chc.ucsb.edu/data/chirps3
- Navigate to the monthly global tifs: https://data.chc.ucsb.edu/products/CHIRPS/v3.0/monthly/global/tifs/
- Download all the February files from 2014 to 2025 (look for files named like
chirps-v3.0.2014.02.tif,chirps-v3.0.2015.02.tif, etc.)
Once downloaded, upload all these files to your Atlas project. You can drag and drop them into the data panel.
Pro tip: Create a folder called "February Rain Data" and organize all your files there using multiselect (hold Shift and click to select multiple files, then move them together into the folder).
Step 3: Preview the Data
Before we start our AI-powered analysis, let's take a quick look at the data. Click on one of the raster files—say, February 2022—and visualize it on your map.
Notice how the southern part of Zanzibar tends to have more rainfall than the northern part? This is exactly the kind of pattern we want to quantify for every hotel location.
Step 4: Start Your Conversation with Navi
Now for the fun part. Click the blue Navi button in the top left corner to open the AI assistant.
Here's your first prompt (copy this exactly or adapt it to your needs):
Hello! I want to find the best place to book a hotel in Zanzibar in February. I have uploaded CHIRPS data for February from 2014-2025. I also want to count how many restaurants that are inside a 15 min walking isochrone for each hotel. Maybe we can start with fetching the polygon for Zanzibar and clip the rasters (maybe a buffer around to see a bit out from the islands border). Can we start by just getting the polygon of Zanzibar without the buffer so we can reuse that in other workflows?
Navi will process this request and return a polygon boundary for Zanzibar from OpenStreetMap.
Perfect! Now we have our study area defined.
Step 5: Fetch Hotels and Restaurants
Next, ask Navi to get all the hotels and restaurants within this polygon. Type:
Can you fetch all hotels and restaurants inside the Zanzibar polygon?
Navi will query OpenStreetMap and return two datasets:
- All hotel locations in Zanzibar
- All restaurant locations in Zanzibar
You should see the data appear on your map. Now we're ready to calculate how many restaurants you can walk to from each hotel in 15 minutes.
Step 6: Calculate Restaurant Accessibility
Now we're ready to answer a key question: how many restaurants can you walk to from each hotel in 15 minutes?
Prompt Navi:
Nice! Lets continue with counting how many restaurants that are inside 15 min walking distance. The output should be the hotels as points with the count on number of restaurants.
Navi will:
- Generate a 15-minute walking isochrone (walkable area) around each hotel
- Count how many restaurants fall within each isochrone
- Add this count as a new attribute to the hotels dataset
This is powerful analysis that would normally require understanding buffer calculations, network analysis, and spatial joins—but Navi handles it all conversationally.
Step 7: Style by Restaurant Count
Let's visualize which hotels have the best restaurant access. Ask Navi:
Can you style the hotels by the number of nearby restaurants? Use the circle radius to show the count.
The results are immediately clear: hotels in Stone Town have the best access to restaurants, with some having 50+ dining options within a 15-minute walk. Hotels on the eastern beaches have fewer nearby restaurants but might offer other benefits (like being on the beach and, as we'll see, having less rain).
Step 8: Analyze Historical Rainfall
Now for the rain analysis. We want to know the average February rainfall for each hotel location based on 11 years of data. Prompt Navi:
Can you now find the mean, std, and p90 for all the hotels for rain score from 2014 to 2025? Also include the same calculations for 2024 only. The score should be merged onto the Hotels_Zanzibar_with_restaurant_count_15min_walk dataset.
Navi might run this as one large workflow or create separate workflows for each raster file—either way, the result is the same. It will:
- Extract rainfall values from each February raster (2014-2025) at each hotel location
- Calculate statistics: mean (average), standard deviation (variability), and 90th percentile (p90) across all years
- Also calculate these stats for 2024 alone for comparison
- Merge these values into your hotels dataset
This takes what could be hours of manual raster analysis and completes it in minutes through AI automation.
Step 9: Style by Rain and Restaurants
Now we can create our final visualization combining both factors. Tell Navi:
Nice! Can we style the dataset? So the color shows the mean rain for 2014-2025 and the radius is number of restaurants?
This is the magic moment. You can now see at a glance:
- Hotels in the north (lighter yellow colors): Less rainfall, perfect if you want guaranteed good weather
- Hotels in Stone Town (larger circles): Maximum restaurant access, great if you want dining variety
- Hotels in the south (darker blue colors): More rainfall, but still viable with good planning
The northeastern coast appears to offer the best of both worlds: relatively low rainfall and moderate restaurant access, plus beach proximity.
Step 10: Add Tourist Points of Interest
Let's make this even more useful by adding other amenities travelers care about. Prompt Navi:
Can you also fetch tourist points of interest within Zanzibar? I'm interested in: aquarium, art_gallery, bakery, bar, beach, cafe, golf_course, hiking_trail, hospital, ice_cream_shop, liquor_store, monument, museum, nature_reserve, park, pharmacy, restaurant, souvenir_shop, supermarket, tennis_court, waterfall, and zoo.
Navi will fetch all these POI types from OpenStreetMap and add them to your map with appropriate styling.
Now you can see not just rain and restaurants, but also beaches, cultural sites, healthcare facilities, and entertainment options near each hotel.
Step 11: Create a Beautiful Interface
Finally, let's make this into something you can easily explore and share. Ask Navi:
Can you create a nice interface for exploring this data? I want to be able to click on hotels to see their details, filter by rain levels, and toggle different POI layers.
Navi will generate an interactive interface with:
- Clickable hotel markers showing detailed statistics (rain data, restaurant count, nearby amenities)
- Filter controls to show only hotels within your rain tolerance
- Layer toggles for different POI types
- A search bar to quickly find specific hotels by name
- A legend explaining the styling
The search widget is especially handy when you're browsing hotels on Booking.com or similar sites—just search for the hotel name in your Atlas interface to instantly see its rain score and restaurant count before you book.
Try clicking around. For example, a hotel in Zanzibar City has:
- Rain score (2015-2025): 30.9
- Restaurants within 15 min walk: 58
Compare that to a hotel in Kizimkazi (southern tip):
- Rain score (2015-2025): 43.2
- Restaurants within 15 min walk: 2
The tradeoffs become immediately clear—Zanzibar City offers walkability and dining options, while the southern beaches are more remote but have that resort feel.
Step 12: Ask Navi for Recommendations
Still can't decide? Just ask Navi to suggest hotels based on your preferences:
Can you suggest some hotels based on the data? I want a balance between good weather and restaurant access.
Navi will analyze the data and give you specific recommendations. Here's what it suggested for me:
If you want the most food/walkability (Stone Town / Zanzibar City)
These come out at the very top for restaurant count (~60–61 restaurants within 15 min walk), but they all have similar rain scores:
- Antonio Garden Hotel
- Spice Palace Hotel
- Aurelia Zanzibar
- Zanzibar Coffee House
- Jambo Guest House
- Pearl Guesthouse
- Riverman Hotel
- Blue House
If you want drier weather (Matemwe / Kaskazini beach area)
These have the lowest rain scores in the dataset, but restaurant access is limited (0–1 nearby)—more remote/resort-style:
- Matemwe Attitude Hotel
- Sunshine Hotel
- Villa Kiva Boutique Hotel Zanzibar
- Infinity Beach Hotel & SPA
- Marafiki Beach Hotel and Spa
- The Seles Boutique Hotel
- Villa Loiki
- Utupoa
This is the power of having your data analyzed by AI—you can ask follow-up questions and get personalized recommendations based on the actual numbers.
What Just Happened?
In about 20 minutes and with a handful of conversational prompts, you've performed sophisticated GIS analysis that would typically require:
- Professional GIS software ($$$)
- Knowledge of raster analysis, vector operations, spatial statistics, and isochrone calculations
- Hours of manual data processing
- Scripting or coding skills to automate the workflow
Instead, you:
- Downloaded free climate data
- Described what you wanted in plain English
- Let AI handle the complex spatial analysis
- Got a beautiful, interactive result you can use to make an informed decision
Making Your Decision
Based on this analysis, here's how you might decide:
If you prioritize weather: Book in the north or northeast (Nungwi, Kendwa, Matemwe areas) where February rainfall averages 30-40% less than the south.
If you prioritize dining and culture: Stone Town offers 3-4x more restaurants within walking distance plus museums, galleries, and historic sites.
If you want balance: The northeastern beach areas (Matemwe, Kiwengwa) offer good weather, moderate restaurant access, and beautiful beaches.
If you're adventurous: The southeast (Jambiani, Paje) has more rain but also has kitesurfing, less tourism, and authentic local experience.
The data doesn't make the decision for you—it empowers you to make an informed choice based on your priorities.
Use Cases Beyond Vacation Planning
While we used vacation planning as a fun example, this same approach works for:
- Real estate investors analyzing property locations by climate risk, amenity access, and demographic patterns
- Business owners evaluating store locations by foot traffic areas, competitor proximity, and transportation accessibility
- Urban planners assessing service coverage by analyzing facility locations against population density and transit networks
- Researchers studying spatial patterns in environmental data, social phenomena, or economic activity
- Emergency managers planning evacuation routes based on population distribution, road networks, and hazard zones
Any decision that involves location, spatial relationships, and multiple data layers can benefit from this conversational AI-powered GIS approach.
Tips for Using Navi Effectively
Start with clear goals - Tell Navi what you're trying to accomplish, not just what operations to perform. "I want to find hotels with low rain and good restaurant access" is better than "clip these rasters."
Break complex tasks into steps - While Navi can handle complex multi-step prompts, breaking things into logical steps gives you more control and helps you verify each stage.
Be specific about data sources - When fetching data from OpenStreetMap, specify exactly what tags or categories you want to avoid getting too many irrelevant results.
Review intermediate results - Check the data and visualizations after each major step to ensure Navi understood your intent correctly.
Iterate on styling - Don't be afraid to ask Navi to restyle visualizations multiple times until they clearly communicate your insights.
Ask for explanations - If you're unsure what Navi did, ask it to explain the workflow, methods, or calculations it used.
The Future of GIS is Conversational
This tutorial demonstrates a fundamental shift in how GIS works. For decades, spatial analysis required specialized training, expensive software, and hours of manual work. Now, AI agents like Navi make sophisticated geospatial analysis accessible through conversation.
You describe your analytical goals in plain English. The AI:
- Understands spatial concepts and operations
- Fetches data from appropriate sources
- Executes complex workflows
- Styles results for clear communication
- Creates interfaces for exploration and sharing
This democratization of GIS means domain experts can directly perform spatial analysis without becoming GIS specialists. A travel blogger can analyze climate data. A small business owner can evaluate locations. A researcher can process satellite imagery. All without taking a single GIS course.
Data-Driven Decisions with Atlas and Navi
When you're making location-based decisions—whether planning a vacation, choosing a business site, or conducting research—having data is just the start. Turning that data into insights requires spatial analysis, and that's where Atlas and Navi work together.
Atlas gives you the platform to work with maps, spatial data, and geographic workflows. Navi gives you an AI assistant that understands geospatial concepts and can execute complex analysis through conversation.
Turn Questions into Analysis
You can:
- Ask spatial questions in plain English and get actionable answers
- Combine multiple datasets (climate, demographics, infrastructure, amenities) into unified analysis
- Generate visualizations that clearly communicate insights to any audience
Also read: AI-Driven Geospatial Solutions: Transforming Location Intelligence
Build Analysis That Scales
Atlas and Navi let you:
- Start with one location and scale to thousands with the same workflow
- Automate analysis that would take hours manually with simple prompts
- Create reusable workflows that can be applied to new data and different regions
That means no more manual calculations for each location, and no more rebuilding analysis from scratch when you want to study a different area.
Analyze Better with AI and Location Intelligence
Whether you're planning a vacation, growing a business, conducting research, or managing operations, Atlas and Navi help you make better location-based decisions through accessible geospatial intelligence.
It's GIS—designed for everyone, powered by AI, accessible through conversation.
Make Better Location Decisions
Location matters for everything from vacation planning to business strategy, but spatial analysis has traditionally required specialized expertise and expensive tools. That's changing.
Atlas gives you professional GIS capabilities in your browser. Navi gives you an AI assistant that understands spatial concepts and can execute complex workflows through conversation.
In this article, we showed how to find the best hotel in Zanzibar using rain data and AI-powered proximity analysis, but that's just one example of what's possible.
From site selection to market analysis, environmental monitoring to urban planning, Atlas and Navi make sophisticated geospatial analysis accessible to anyone with data and questions. All from your browser. No GIS expertise needed.
So whether you're planning your next vacation, choosing a business location, or conducting spatial research, Atlas and Navi help you move from "where should I go?" to "here's exactly where, and here's the data to prove it" faster.
Sign up for free to try Navi in open beta, or book a walkthrough to see how Atlas can transform your location-based decisions.
