pandas is the table library. A pandas DataFrame is a spreadsheet you drive with code: rows and columns, a header, numbers read as numbers. This page maps a handful of pandas moves onto the Excel moves you already know — a table, AutoSum, AutoFilter, a PivotTable, and a lookup. We keep it shallow this week. Project 2 goes deeper. Read each Predict card first, then press Shift+Enter (or click Run) to execute the cell. This is real Python, running in your browser.
A short table of AR invoices. There is no file to load — the rows below live in a string inside the page, and the first cell reads them into a DataFrame. Everything after that works on this one table.
io.StringIO, so you can focus on the pandas moves without a file getting
in the way. The shape is exactly what pd.read_csv("invoices.csv") would give you.
Run the cells top to bottom. Each one does in pandas what you used to do with a click in Excel. The last cell is yours to experiment in.
That is the shallow end of pandas: read a table, total a column, filter rows, group and aggregate, and join two tables. Every one of those is a spreadsheet move you have already done by hand — the difference is that the pandas version is written once and runs the same way on seven rows or seven hundred thousand.
read_csv, a column with .sum(), a boolean-mask filter,
groupby, and a merge — are the ones you will reach for first. Get comfortable with the
Excel-to-pandas mapping now, and Project 2 is reading and grouping, not learning the syntax cold.