How to Use Quadratic for AI Spreadsheet Analysis

Quadratic is a web-based AI spreadsheet that runs Python, SQL, and JavaScript directly inside cells. You connect your data, describe what you want in plain English, and the AI writes the code and runs it for you. No local setup, no separate IDE, no pasting results back into a spreadsheet.

What Makes Quadratic Different

Most spreadsheet tools treat code as an external step. You run an analysis in Python or SQL, then paste the results back into Excel or Google Sheets. Quadratic merges the two environments: each cell can contain a formula, a Python snippet, a SQL query, or an AI-generated analysis block, all within the same infinite canvas.

The platform is built on Rust, WebAssembly, and WebGL, which lets it handle larger datasets and execute code faster than a typical browser-based spreadsheet. Under the hood, Quadratic uses Claude 3.7 Thinking, Claude 3.5 Sonnet, and GPT-4o to power its AI code generation features.

This architecture matters because the AI does not produce a text answer. It generates working, inspectable code that runs in your sheet and can be reviewed, modified, and rerun at any time. You are not trusting a black box; the logic is always visible.

The result is a tool that sits between a traditional spreadsheet and a Jupyter notebook without requiring the setup of either. For analysts who spend time copying results between Python scripts and Excel files, this is the gap Quadratic fills.

Setting Up Your First Sheet

Quadratic runs entirely in the browser. Go to app.quadratichq.com, sign up with a Google account or email, and you land in a blank sheet. There is nothing to install or configure locally.

The interface looks familiar: rows, columns, a formula bar at the top. The key addition is a right-side panel that switches between three modes: AI (for natural language requests), Code (for directly writing Python, SQL, or JavaScript), and Connections (for linking to databases).

A free plan is available with limits on AI usage. Paid plans unlock higher AI call limits, additional database connection types, and team collaboration features. For individual analysts running exploratory analyses, the free plan is enough to evaluate the tool thoroughly before committing.

Importing and Connecting Your Data

To analyze data you already have, drag a CSV or Excel file directly into the sheet. Quadratic reads the file and places the data starting at your selected cell. Column headers are preserved and visible immediately.

For live data, Quadratic connects to PostgreSQL, MySQL, Snowflake, and other databases. Open the Connections panel, enter your credentials, and write a SQL query in a code cell. Quadratic executes the query against your database and returns results as a spreadsheet table that refreshes each time you run the cell.

This distinction matters for operational teams. A CSV import is a snapshot; a database connection keeps the analysis synchronized with current data. Finance teams running monthly close reports, for example, can connect directly to their reporting database and refresh the analysis in one click rather than exporting a new CSV each time.

Quadratic also supports connections to third-party APIs, which means you can pull in external datasets alongside internal data in the same sheet.

Using AI to Analyze Your Data

Click any empty cell and open the AI panel. Type a question or request in plain English. The AI reads the context of your sheet, including column names, visible data ranges, and any code in adjacent cells, then generates Python or SQL code that performs the requested analysis.

Examples of prompts that work well:

• "Summarize the key trends in this sales data by region and quarter"
• "Calculate month-over-month growth rate for each product category"
• "Run a linear regression on column D against column F and explain the R-squared value"
• "Find all rows where revenue is more than two standard deviations below the rolling 90-day average"

The generated code appears in the cell editor before execution. You can review it, edit individual lines, and run it. If the output is not what you expected, type a follow-up prompt to refine the code further.

This workflow is more reliable than asking a general-purpose AI chatbot for analysis because Quadratic has direct access to your actual data structure. The AI knows your column names, data types, and row count, so it generates code specific to your dataset rather than a generic template.

Building Charts with AI

To create a visualization, select a data range and open the AI panel. Describe the chart you want: "Bar chart comparing Q1 sales by product, sorted descending, with a horizontal reference line at 50,000." Quadratic generates Python code using matplotlib or plotly and renders the chart directly in the sheet.

Charts are positioned as embedded objects on the infinite canvas. You can drag them next to the data they reference, resize them, and they update automatically when the underlying code cell reruns.

For standard charts without custom requirements, Quadratic also has formula-based chart options similar to Google Sheets, which do not require AI or code. The AI chart path is most useful when you need conditional formatting, multi-series overlays, custom annotations, or statistical overlays like trend lines or confidence intervals.

Writing Python and SQL Directly

For users comfortable with code, Quadratic lets you write Python or SQL in any cell without the AI. Python cells use a sandboxed environment with pandas, numpy, matplotlib, and plotly available by default. A cell outputs its result (a value, a DataFrame, or a chart) to the sheet range you specify.

SQL cells execute directly against your connected database and return results as a table in the sheet. You can reference the output of a SQL cell in a Python cell downstream, enabling a full data transformation pipeline inside a single file.

This is useful for analysts who want to prototype a data pipeline without setting up a local Python environment or a Jupyter notebook server. The sheet becomes the analysis artifact, with every step visible and reproducible by anyone who opens the file.

A Practical Workflow Example

A sales manager receives a monthly CSV export from their CRM with columns for deal owner, region, pipeline stage, and close value. The usual workflow involves Excel pivot tables, VLOOKUP formulas, and manual chart building that takes the better part of an afternoon.

In Quadratic, the same analysis takes these steps:

1. Drag the CSV into the sheet. Data loads immediately with headers intact.
2. Open the AI panel and type: "Build a summary showing total close value by region and sales rep. Highlight the top three performers."
3. Review the generated pandas code, click Run.
4. Type: "Create a bar chart of close value by region, sorted highest to lowest."
5. Review the chart output and adjust axis labels if needed.

The whole process takes under five minutes. The sheet retains each step as reproducible code cells. When the CRM exports updated data next month, the manager re-uploads the CSV and reruns the cells with no rebuild required.

If you want to skip the export-and-upload step and connect a data source directly for automated analysis from a single prompt, VSLZ AI handles that from a file upload with no configuration required.

Limitations to Know

Quadratic's free plan limits the number of AI calls per month, which can be restrictive for daily analysis work. Heavy Python computations or queries against very large database tables may encounter memory constraints in the browser environment.

The platform does not support VBA macros, and complex Excel formulas using INDIRECT, OFFSET, or older array syntax may need to be rewritten. Migrating an existing complex Excel model is rarely straightforward; Quadratic is best suited for new analyses rather than direct conversions of legacy spreadsheets.

For teams already deeply embedded in Excel workflows, the migration overhead matters. Quadratic works best when adopted for new projects or analytical work that would otherwise require a separate Python environment.