•
Excel Copilot and Python in Excel (=PY) are useful at work. Tonight export one sheet, plot one series in a real .py file, and save a PNG you can version-control.
You already know how to make a chart in Excel. That skill is not obsolete. Microsoft has put Python inside the spreadsheet: you can type =PY into a cell, and Copilot in Excel can help you insert those formulas. The after-work question is not “should I abandon Excel?” It is: will tonight’s chart live only in a workbook, or also as a file you can re-run, commit, and send as a PNG?
This is for working professionals who want to stay competitive — not a bootcamp, not a career switch. One sheet, one series, one PNG. That is a complete evening.
Python in Excel lets you write Python in a worksheet cell. Microsoft’s introduction is here: Introduction to Python in Excel. You type Python into a cell (=PY from the formula bar, or Formulas → Insert Python). Calculations run in the Microsoft cloud, not on your laptop’s Python install. You need internet. You do not need a local Python for the feature. If you already have Python on the machine, your venv and pip packages do not apply inside those cells.
The cloud runtime ships a core Anaconda set: pandas, matplotlib, seaborn, statsmodels, and NumPy by default; scikit-learn (sklearn) is available to import. You can clean a table, fit a small model, and plot without leaving the workbook. Colleagues still open Excel.
Copilot in Excel now has a Python skill. In What’s New in Excel (August 2026) (25 August 2026), Microsoft lists “Copilot in Excel: Python skill (Insiders)”: analyse and visualise with Python in Excel, create editable, refreshable Python formulas, and start with @python-in-excel. Chart and PivotTable skills landed in the same update. If your daytime job is a shared workbook, this is a legitimate way to go deeper without a separate IDE.
Do not treat “Python in Excel” as fake Python. The language is Python; the libraries are ones analysts actually use. The limitation is the container: cloud execution, a curated package list, no arbitrary pip install, no .py on disk, no git history of the script, no scheduler, no run on a machine without Excel.
A practical split:
=PY, charts, PivotTables. You are not “behind” because you did that.pip, a folder, a .py file) if you want files you own, version control, extra packages, and scripts that run outside Excel.Python in Excel is useful and limited. Learning that only happens inside =PY will not teach git, pin versions, or generate a PNG on a machine that never opens Microsoft 365. Tonight still uses Excel as the source of numbers. It parks the chart in a script.
You should be able to run a .py file and install packages. If lists, functions, and imports still feel new, start with Introduction to Python Programming — 19 published lessons, written for evenings. Check the course page for the current enrollment window; this article will not freeze a date in the title. You do not need pandas fluency tonight — read_csv is enough — and you do not need machine learning.
pip install pandas matplotlib
Keep a simple range: category or date in the first column, a number in the second. No merged cells, no decorative headers, no footnotes in column Z. File → Save As → CSV.
Example monthly.csv:
month,amount
2026-01,1200
2026-02,1500
2026-03,1100
2026-04,1800
If the real sheet has ten metric columns, pick one series. You could plot the same series inside =PY; the point of this evening is a PNG beside a script.
import pandas as pd
import matplotlib.pyplot as plt
df = pd.read_csv("monthly.csv")
print(df)
plt.figure(figsize=(8, 4.5))
plt.plot(df["month"], df["amount"], marker="o")
plt.title("Monthly amount")
plt.xlabel("Month")
plt.ylabel("Amount")
plt.tight_layout()
plt.savefig("monthly_amount.png", dpi=120)
plt.show()
Load, plot, save. savefig matters more than show() if you will paste the figure into a slide. The PNG is the deliverable.
For categories that are not time, a bar chart is usually clearer: plt.bar(df["month"], df["amount"]).
If months plot in the wrong order, they may still be strings:
df["month"] = pd.to_datetime(df["month"])
df = df.sort_values("month")
If encoding looks wrong (common with Excel on Windows), try encoding="utf-8-sig". Check df.tail() so a total row does not become a data point.
plt.xticks(rotation=45) or a coarser summary.Once the table is clean, pandas can plot in one call:
df.plot(x="month", y="amount", kind="line", marker="o", title="Monthly amount")
plt.tight_layout()
plt.savefig("monthly_amount.png", dpi=120)
Use that when you already think in DataFrames. Use plt.plot when you want each step visible. Both call matplotlib. Similar pandas code can live in a =PY cell; you cannot savefig next to a git repo without leaving Excel.
Save monthly.csv next to the script:
import pandas as pd
import matplotlib.pyplot as plt
df = pd.read_csv("monthly.csv")
plt.figure(figsize=(8, 4.5))
plt.bar(df["month"], df["amount"])
plt.title("Monthly amount")
plt.xlabel("Month")
plt.ylabel("Amount")
plt.tight_layout()
plt.savefig("monthly_bar.png", dpi=120)
print("Saved monthly_bar.png")
If the image matches the four numbers, you have crossed from “I use Excel charts” (or Copilot charts) to “I can emit a chart from a file.” Next month, replace the CSV. Do not rebuild the figure by clicking unless you are annotating it.
=PY result if that is where you cleaned it).head() and tail().monthly_amount.png), not figure1.png.If you later groupby, plot the summary, not every invoice. Colleagues will still want Excel: send the PNG plus the CSV, and the script to anyone who will reproduce it. You have moved the chart off a click-path — including a Copilot prompt you might not remember — into a folder you control.
Default matplotlib is enough. Prefer a sentence title (“April amount higher than Q1 months”). Start bar-chart y-axes at zero. Skip dual axes and 3-D bars. Five minutes on title, units, and tick rotation beats a new library.
Do not start with dashboards. Do not skip local Python because =PY already imported pandas. Do not assume Copilot’s Python skill replaces a script you can schedule. This is not a get-rich plan. The competitive move for someone already employed is a repeatable evening habit: one honest figure, one file, one PNG you can find on Friday.
If you still need the interpreter, pip, and the first programs that use a library, the live path is Introduction to Python Programming — 19 published lessons, built for after work. Then come back to this chart with a .py file that is yours.