What You’ll Learn in This Guide
1. Create a new map
You can create a new map in Neural Earth Studio in two different ways — either directly from the Maps page or while browsing the Data Catalog. Your new map automatically saves to your account, so you can return to it anytime from the Maps page.
From the Maps page
Sign in to your Neural Earth Studio account.
Navigate to the Maps page — this page lists all your existing maps.
Click + New Map in the top-right corner.
A new map will open in your current browser tab, ready for setup.
From the Data Catalog
Browse or search for the dataset you want to use in the Data Catalog.
On the dataset page, click + Add to map in the top-right corner.
From the dropdown, select Create new map.
Neural Earth Studio will automatically create a new map with the selected dataset already loaded and visible in the left-hand panel.
💡 Pro Tip: This method is perfect when you want to start a new project directly from a dataset in the Data Catalog without creating a blank map first.
2. Explore your map workspace
Once your new map opens, you can start customizing it for your project needs. Inside the map, you can:
Bring your own data: Upload files like CSV, GeoJSON, or Arrow for custom datasets.
Add data from the Data Catalog: Browse preloaded datasets available in Neural Earth Studio.
Create data layers: Organize multiple datasets into layers for visualization and filtering.
Interact with Gaia, your AI Assistant: Use natural language to add datasets, run analysis, or automate styling.
Customize visualization: Adjust basemaps, colors, and layer opacity for clarity.
💡 Pro Tip: Start by uploading one dataset to familiarize yourself with the workspace before adding multiple layers.
3. Add data to your map
After creating a new map, click + Add Data in the left-hand column.
A pop-up titled “Add data to this map” will appear with two options:
Load Files — upload local data directly.
Import Tileset — connect to remote datasets via URL (vector, raster, or WMS).
How to upload files
Drag and drop your file(s) into the upload window, or click Browse to select files.
Supported file types: CSV, JSON, GeoJSON, Arrow, Parquet, and saved Map JSON.
Maximum upload size: 250 MB.
Once uploaded, the dataset will automatically load into your map and appear in the left-hand column with its name and row count.
How to add data from the Data Catalog
You can add datasets from the Neural Earth Studio Data Catalog directly into your map in two ways — either from inside an existing map or directly from the catalog page.
From within a new or existing map
Inside your map, click Add data from catalog in the left-hand column.
A new browser tab will open displaying the Data Catalog.
Browse or search for the dataset you want to use and click on it.
On the dataset page, click + Add to map in the top-right corner.
From the dropdown menu, select Add to an existing map.
In the pop-up, use the search bar to find your target map by name.
Select the map and click Add to map to confirm.
You’ll receive a platform notification confirming the dataset has been added.
Return to your map to verify that the dataset now appears in the left-hand panel.
From the Data Catalog page
Visit the Data Catalog directly from the main navigation or a new browser tab.
Browse or search for the dataset you want to use and click on it.
On the dataset page, click + Add to map in the top-right corner.
From the dropdown menu, select Create new map.
Neural Earth Studio will automatically create a new map with the selected dataset already added and visible in the left-hand panel.
4. Upload vector, raster, or WMS tiles
To import external map layers, select Import Tileset in the same Add Data pop-up.
Choose the Tileset Type: Vector Tile, Raster Tile, or WMS.
Enter a valid URL for your tileset.
Must contain at least one dot (e.g.,
example.com).Allowed characters: letters, numbers,
-,_,/,{},.,?,=,&.If missing a scheme,
https://is assumed.
Provide a Name for your dataset (e.g., “NOAA Flood Zones” or “Population data”).
Optionally, include a Tileset Metadata URL (recommended).
Click Load tileset to import.
💡 Pro Tip: For WMS uploads, ensure the endpoint is a valid WMS service URL. The “Load tileset” button remains disabled until the URL passes validation.
5. Add multiple datasets
To add more data to your map:
Click Add More Data in the top-right corner of your workspace.
Choose from:
Upload local files (CSV, GeoJSON, JSON).
Paste a URL to load remote datasets directly (bypasses the 250 MB limit).
Select from the Data Catalog for preloaded Neural Earth datasets.
Repeat as needed — you can add multiple datasets to one map.
💡 Pro Tip: Avoid overloading your map with large datasets simultaneously to maintain optimal performance.
6. File format guidelines and limits
Neural Earth Studio supports a variety of common geospatial and tabular file types for flexible data ingestion. Each format has specific capabilities and size considerations depending on your data source.
The following table outlines all supported file formats, their descriptions, and important usage notes.
File Type | Description | Supported | Notes |
CSV | Tabular data with coordinates or location columns | ✅ | Max 250 MB |
JSON / GeoJSON | Structured geographic data | ✅ | Auto-parsed and rendered |
Arrow / Parquet | High-performance binary formats | ✅ | Ideal for large datasets |
Map JSON | Saved Neural Earth Studio map | ✅ | Restores prior configurations |
Vector Tile | External tile server | ✅ | Must be valid URL |
Raster Tile | Cloud-optimized GeoTIFF tiles | ✅ | Limited STAC support |
WMS | OGC Web Map Service endpoint | ✅ | Must be valid service URL |
7. Supported coordinate systems and projections
Neural Earth Studio supports Web Mercator (EPSG: 3857 – WGS 84) only. All geometry coordinates must:
Use the WGS 84 datum.
Represent longitude and latitude in decimal degrees.
💡 Pro Tip: Ensure your dataset’s CRS matches EPSG: 3857 to avoid misaligned features on the map.
8. Auto layer detection and data type recognition
When uploading CSV or GeoJSON files, Neural Earth Studio automatically detects data types and creates appropriate layers.
Data type detection
Column types are identified automatically (boolean, date, integer, real, string, timestamp, or geometry). If a column mixes data types, it defaults to string for safety.
Examples:
True / False→ Boolean2019-01-01→ Date{"type":"Polygon","coordinates":[…]}→ GeoJSON geometry
Auto layer creation
Layers are auto-generated if column names follow these conventions:
Detected Pattern | Auto-Created Layer |
| Point Layer |
Two sets of coordinates (e.g., origin_lat/origin_lon and dest_lat/dest_lon) | Arc Layer |
Column named | Icon Layer |
Column named | H3 Layer |
Column containing GeoJSON or WKT geometry | Polygon / GeoJSON Layer |
💡 Pro Tip: Studio also reads style declarations inside GeoJSON files (e.g., lineColor, lineWidth, fillColor, radius) to auto-style your features.