# Data Science & Machine Learning

Built with technologies like [Apache Arrow](https://docs.spice.ai/~/changes/fl1GUvkHE2461izfYRLa/api/sql-query-api/apache-arrow-flight-api), Spice is designed from the ground-up for data-driven apps, data science and machine learning.

Get started with the following guide.

### 1. Find the data you need

Spice includes a growing set of web3 data, including blockchain data, cryptocurrency and token prices, ENS domains, and more.

<figure><img src="https://692214342-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-Mc71sWdWMqp1e_rsTZE-129938315%2Fuploads%2Flv5l5x7TTbUc4hVxQfoD%2Fimage.png?alt=media&#x26;token=c62cba92-28b4-4fdb-8e0f-7917eb0b6146" alt="" width="375"><figcaption><p>The dataset reference in the SQL Playground.</p></figcaption></figure>

Explore the full list of [**datasets**](https://docs.spice.ai/~/changes/fl1GUvkHE2461izfYRLa/getting-started/datasets) for an overview and see [**SQL Query Tables**](https://docs.spice.ai/~/changes/fl1GUvkHE2461izfYRLa/data-tables) for schemas and details.

### 2. Query datasets using SQL

Querying datasets is as easy as querying data from any SQL database. Try SQL from your browser at [**Spice.xyz**](https://spice.xyz/).

<figure><img src="https://692214342-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-Mc71sWdWMqp1e_rsTZE-129938315%2Fuploads%2FRlNA7KLFMmDmvMmHzyr4%2Fimage.png?alt=media&#x26;token=9ac416ea-7b6c-4a4d-b29e-f6c860af6b80" alt="" width="563"><figcaption><p>Querying datasets with SQL in the Playground.</p></figcaption></figure>

Reference [**SQL best practices**](https://docs.spice.xyz/best-practices) for tips and the best query performance.

### 3. Refer to the SQL reference

Spice uses an [**Apache Calcite**](https://calcite.apache.org) based query engine, which supports **ANSI SQL** with additional SQL **dialect**.&#x20;

<figure><img src="https://692214342-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-Mc71sWdWMqp1e_rsTZE-129938315%2Fuploads%2FiWRWIKDLf3DpYnGNUQoc%2Fimage.png?alt=media&#x26;token=bf8d2c66-4ec0-4bb0-8a04-704459e173c0" alt=""><figcaption><p>The SQL Reference at docs.spice.xyz</p></figcaption></figure>

Refer to the [**SQL reference**](https://docs.spice.ai/~/changes/fl1GUvkHE2461izfYRLa/reference/sql-reference) for dialect specific data types, functions, and commands. SQL keywords are also indexed in search for quick lookup.

### 4. Use the Python SDK

Import Spice data in notebooks like Kaggle, Jupyter Notebooks, Google Colab, Anaconda Notebook, etc. with [**3 lines of Python code**](https://docs.spice.xyz/sdks/python-sdk#usage).

<figure><img src="https://692214342-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F-Mc71sWdWMqp1e_rsTZE-129938315%2Fuploads%2FkTQ8wowoPFksabkl6gUK%2Fimage.png?alt=media&#x26;token=850c8bd4-a83c-4dda-9da5-df0e5430be44" alt=""><figcaption><p>Querying Spice from Python in 3 lines of code.</p></figcaption></figure>

### 5. Use familiar data science tools and libraries

Easily use Python libraries like **numpy, pandas, pyplot, sklearn, xgboost, and more** to perform **exploratory data analysis (EDA),** create articulate **visualizations**, and build dynamic **machine learning models.**

See sample Kaggle Notebooks on the next page --> <br>