Data & spreadsheets · course 43 of 105

Your Spreadsheet in Code with pandas

Move your spreadsheet work into Python: build a table, pick rows and columns with .loc and .iloc, filter, sort, fill missing values, group, merge and count with pandas.

4 modules

Try a question

Not on sale yet. When sales open: $4.99 USD for this course, paid once, or every course for $9.99 USD a month. Prices

Runs in Claude (claude.ai, the Claude apps or Claude Code) with your own Claude account. A course opens there as soon as your payment goes through; nothing is shipped.

Preview

What it teaches

The first pandas an analyst needs when spreadsheet work moves into Python. What a Series and a DataFrame are (quoted from the pandas 3.0.6 documentation); building a table from a dict or a list of rows; its shape; picking columns and rows by label (.loc, where a slice includes its end) and by position (.iloc, where it doesn’t); filtering rows with True/False conditions (each condition in brackets, & for and, | for or); sorting while the row labels travel with their rows; adding a column; missing values (NaN, count() vs len(), fillna, dropna, why whole numbers turn into decimals); group-by with sum, mean, count, min, max and reset_index(); merging two tables (inner vs left, a repeated key repeats the row); and value_counts(). Two kinds of question: "what does this code print?" and "write one line that makes this table" (graded on the table your line makes). The AI asks 3 questions, places you, shows a short card per module, then gives you one question at a time.

Not an install guide, not a data-cleaning course, and it can’t debug your own file.

How long

33 questions plus 10 concept questions. About 12-30 rounds for someone who already knows some pandas, up to about 60-80 for a beginner, plus a 12-item final: 2-4 sessions (estimate, not timed with a real student). The tutor gives a one-line save code at the end of each session.

The modules

4 modules, drilled one item at a time.

M0
The words
The words: Series, DataFrame, .loc against .iloc and which slice end is included, missing values, group by, merge, value_counts()
M1
DataFrames: build, shape and select
Build a DataFrame, its shape, and pick columns and rows by label (.loc) and by position (.iloc)
M2
Filter, sort, new columns and missing values
Filter rows, sort, add a column, and handle missing values
M3
Group, merge and count
Group-by with sum, mean and count, merging two tables, and counting values with value_counts()

Before it teaches, it asks

1
What do you do with spreadsheets or tables now (reports, finance, surveys, nothing yet), and can you run Python with pandas on your computer right now (fine to say no)?
2
What does this print? If it stops with an error, name the error.df = pd.DataFrame({"name": ["Ana", "Bo", "Cy"], "age": [31, 25, 40]})print(df.shape)
3
What does this print? If it stops with an error, name the error.df = pd.DataFrame({"x": [4, None, 8, None, 6]})print(len(df), df["x"].count())

What’s exact, and what isn’t

Show

Straight from the course’s own notes.

Keys come from running the code: Every M1, M2 and M3 key, every number in the working lines and every table or value shown on the M1-M3 cards was produced by running the code on pandas 3.0.6 (Python 3.14.3, numpy 2.5.4), in a fresh namespace with every warning turned into an error, and again in a fresh interpreter process (133 snippets, no difference). 12 of the short keys were also written by hand first and must equal the run.
Wrong answers are proven: All 40 listed wrong answers. For the 26 on “what does this print?” items, the slip is written as a one-edit change of the item’s code, and the answer key shows what that code really prints. For the 14 on “write one line” items, the wrong line was run and the answer key shows what it really prints or raises. All 12 “Also right” lines (other lines that are accepted as correct) were run and make the check print the key (one of them differs only in the Series footer line, which the tutor doesn’t grade).
M0 is quotes, not computed: M0 and its Answers are built from 21 quotes. 4 more quotes sit on the M2 and M3 cards (the bracket rule for & and |, isin, the group sort, group keys as the index) and 1 more explains why no value_counts item has tied counts. All 26 come from 7 pandas documentation pages (headed “pandas 3.0.6 documentation”, read 2026-10-13), each checked word for word on a saved copy by a script. The script does not re-fetch the live pages. The M0 Answers are short wordings of those quotes.
Plain-words rules are the course's own: The card rules are written around the quotes and the run examples. Each behaviour they claim is either shown by a run example on the card or checked by a run that the card does not show.
Judgment: If your line isn’t one the answer key lists, the AI traces it on the setup and shows a Check: line. It can only run pandas if your app has a code tool, and it says so. If you can run pandas, what your output shows decides what your line does; it never changes a key.
Honest limits: One pandas version, 3.0.6. Other versions may print some things differently; that was not tested. The tables are small (at most 5 rows and 4 columns; the three value_counts items have 6 to 8 rows of one column), so a narrow window should not wrap them. The course drills pandas on its own small tables and can’t judge your data, your files or your error messages.
Not in it yet: Reading and saving files (CSV, Excel), data types and conversions, dates, text columns, apply, pivot_table, concat, grouping by several columns, the other merge types, plotting.