> ## Documentation Index
> Fetch the complete documentation index at: https://docs.delphina.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Try the Sandbox

> Public datasets you can explore in Delphina without connecting your own warehouse

The sandbox lets you try Delphina against real, substantial data before you connect a warehouse of your own. Each dataset below is already loaded and documented in a knowledge base, so you can open a chat and start asking questions immediately — the same way you would against your own data.

## Getting access

Anyone can use the sandbox. To get in, sign in at [analytics.delphina.ai](https://analytics.delphina.ai); there's no separate sign-up form. If your email domain isn't linked to a Delphina organization, your new account joins the public sandbox organization, where you can chat with the agent and build Data Apps against these datasets. The sandbox's knowledge base is read-only.

If your company already uses Delphina, signing in with your work email takes you to your company's organization instead.

## Datasets

These pages are reference material for the data itself: what each dataset covers, how it's structured, how fresh it is, and the kinds of questions it answers well. Read one before you start exploring and you'll get better answers faster.

> [card] [FIFA World Cup 2026](https://docs.delphina.ai/sandbox/fifa-world-cup-2026.md)
>
> All 22 earlier World Cup editions, 49,000+ international matches, and
> the full 48-team 2026 tournament with match stats and betting odds.

> [card] [NCAA Men's Basketball](https://docs.delphina.ai/sandbox/ncaa-basketball.md)
>
> 11 seasons across 368 D1 programs, through 2025-26 — box scores, advanced
> analytics, and 20 million play-by-play events.

> [card] [Anthropic Economic Index](https://docs.delphina.ai/sandbox/anthropic-economic-index.md)
>
> Occupational AI adoption, autonomy levels, and geographic distribution,
> linked to O\*NET task classifications.

> [card] [NYC Taxi & Rideshare](https://docs.delphina.ai/sandbox/nyc-taxi-rideshare.md)
>
> 2.2 billion+ trip records across yellow taxis, green taxis, Uber, and
> Lyft, published by the NYC TLC.

## Getting the most out of it

* **Ask in plain English.** Start broad ("what were the biggest upsets?") and let Delphina show you its SQL, then refine.
* **Look at the SQL.** Every answer shows the query it ran, which is the fastest way to learn how the dataset is shaped.
* **Try a Data App.** Once an analysis looks right, ask Delphina to turn it into a [Data App](https://docs.delphina.ai/analytics-agent/data-apps.md) — a live dashboard you can edit and share.
* **Check the critic.** Each answer is reviewed by a [critic agent](https://docs.delphina.ai/analytics-agent/critic.md) whose verdict appears beneath it.

When you're ready to point Delphina at your own data, see [Warehouse Connections](https://docs.delphina.ai/administration/warehouse-connections).


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.