How to Use Python in Excel

Python in Excel lets you run Python code inside spreadsheet cells using Microsoft's cloud infrastructure. No local Python installation is required. You get access to pandas, matplotlib, NumPy, and other Anaconda libraries through the =PY() formula. To start, open Excel for Microsoft 365 on Windows or Mac, navigate to the Formulas tab, and select Insert Python, or type =PY( into any cell.

What Python in Excel Can Do

Python in Excel connects your spreadsheet data to a full Python environment running in Microsoft's cloud. When you enter a Python formula, Excel sends the data and code to an Anaconda-managed container, executes it, and returns the result as a value, a DataFrame object, or a rendered chart.

This closes the gap between what Excel handles natively and what data analysts typically need Python for: cleaning messy datasets, running statistical models, generating correlation matrices, or producing publication-quality charts. The environment includes the complete Anaconda distribution, which covers over 1,500 scientific Python packages according to Microsoft's documentation.

For teams that already work in Excel, this means significantly less friction. An analyst can apply a pandas groupby aggregation, run a logistic regression with scikit-learn, or plot a seaborn heatmap without leaving the spreadsheet they were already working in. This is particularly useful for quarterly reporting workflows where the final deliverable is an Excel file and switching tools mid-analysis creates unnecessary overhead.

Requirements Before You Start

Python in Excel is available on Microsoft 365 Personal, Family, Business, and Enterprise plans. It is not included with standalone Office licenses such as Office 2021 or Office 2019. You need an active Microsoft 365 subscription.

As of 2026, the feature runs on Excel for Windows, Excel for Mac, and Excel for the web. It is not available on mobile devices including iPad, iPhone, or Android.

To verify your eligibility: open Excel, go to the Formulas tab, and look for the Insert Python button in the ribbon. If it is absent, your Office installation may need updating. Update via File, Account, then Update Options to ensure you are running version 2406 or later.

Python in Excel requires an active internet connection because all execution happens on Microsoft's servers. If you work offline or on a corporate network with outbound restrictions, Python cells will not recalculate.

How to Write Your First Python Formula

Select an empty cell. Type =PY( and press Tab to confirm. A dedicated code editor appears below the formula bar, giving you a multi-line Python input area.

To reference existing spreadsheet data inside Python, use the xl() function. For example, if your sales data occupies the range A1:C500, reference it as follows:

import pandas as pd
df = xl("A1:C500", headers=True)
df.describe()

This returns a summary statistics DataFrame directly into the cell. By default, Excel displays a collapsed card icon. Click it to expand the DataFrame as a table into adjacent cells.

Set the output type using the formula bar dropdown: Card (the default), Array (spills values into adjacent cells), or Plot (for charts). Most practical use cases start with Array to push results directly into your spreadsheet layout.

Practical Examples for Business Analysis

Summarize sales by region:

df = xl("A1:D1000", headers=True)
df.groupby("Region")["Revenue"].sum().reset_index()

This collapses a thousand-row table into a grouped summary. No pivot table configuration required. Useful when you need to share a clean summary with stakeholders who are not familiar with pivot table mechanics.

Find correlation between two columns:

df = xl("B1:C500", headers=True)
df.corr()

The output is a correlation matrix. Useful for checking whether ad spend actually tracks with sales, or whether two reported metrics are measuring essentially the same thing.

Flag statistical outliers:

df = xl("A1:B300", headers=True)
mean = df["Sales"].mean()
std = df["Sales"].std()
df["outlier"] = df["Sales"] > mean + 2 * std
df

This appends an outlier flag column to your data. Spill the result into an adjacent range and apply native Excel conditional formatting on top of it for immediate visual review.

Working with Charts and Visualizations

Python in Excel renders matplotlib and seaborn plots directly inside the spreadsheet. Change the formula output mode to Plot to display charts inline.

A histogram of customer order values:

import matplotlib.pyplot as plt
df = xl("A1:B1000", headers=True)
fig, ax = plt.subplots()
ax.hist(df["Order Value"], bins=30, color="#0055ff")
ax.set_xlabel("Order Value ($)")
ax.set_ylabel("Count")
plt.tight_layout()
fig

The chart appears as an embedded image in the cell. Resize it like any Excel chart. It recalculates automatically when the underlying data changes.

One difference from native Excel charts: Python plots do not link to slicers or pivot table filters. They recalculate when source data changes, but they do not respond to interactive filter controls. For dashboards where end users apply filters, use native Excel chart types for interactive views and reserve Python charts for analytical outputs.

Limitations to Know Before Relying on It

No external file imports. The common pandas functions pandas.read_csv() and pandas.read_excel() are blocked inside Python in Excel. You cannot pull data from a file path or external URL. All data must come through Power Query or by referencing existing Excel table ranges via xl(). This is the most common friction point for users migrating from a standard Python workflow.

Cloud execution adds latency. Each =PY() formula recalculates by sending code to Microsoft's servers. On large datasets or complex operations, this introduces a delay of several seconds. For spreadsheets with heavy computation across many Python cells, switching to Manual calculation mode reduces recalculation overhead. Access this via Formulas, Calculation Options, Manual.

No cross-cell state. Variables defined in one =PY() cell are not available in another. Each formula runs in isolation. To share a computed DataFrame across multiple output cells, structure your analysis to flow from a single parent formula, or use Python's xlwings-style approach of returning arrays that downstream cells reference.

Row limits apply. Excel's 1,048,576-row limit applies to the source ranges you pass to xl(). Python itself can handle larger data, but your input cannot exceed the spreadsheet's capacity.

If your workflow requires connecting live databases, running multi-step pipelines, or analyzing data too large to load into a spreadsheet, VSLZ AI handles this from a plain-English prompt after a file upload, without any formula syntax or environment configuration.

What This Means for Everyday Analysts

Python in Excel is a practical upgrade for anyone who already works in Excel and wants more analytical power without switching tools. A survey conducted by Microsoft in late 2025 found that over 60% of Excel users who tried Python in Excel used it weekly within the first month, primarily for data cleaning and chart generation tasks that previously required exporting to a separate Python environment.

The setup is minimal: an active Microsoft 365 subscription, Excel version 2406 or later, and an internet connection. The core workflow of referencing data with xl(), applying pandas or scikit-learn operations, and returning results as a table or chart takes about five minutes to learn.

The main constraints to plan around are the block on external file imports, the latency on large datasets, and the lack of cross-cell variable persistence. For analysis that fits within a single connected session and uses data already in the spreadsheet, Python in Excel removes most of the traditional barrier between Excel and Python workflows.