Python vocabulary: Journeyman level

168 Python terms at the Journeyman level, each with a short definition and a link to its full page.

Updated

Study these as flash cards

__init__ (Constructor)
Runs automatically when an object is created and sets up its starting data.
__init__.py
A file that marks a folder as a Python package.
__str__ and __repr__
Control how an object looks when printed (__str__) and when debugging (__repr__).
agg()
A pandas method that works out several summaries for each group at once, such as a sum, a mean and a count, and hands them back as one table.
annotate()
A matplotlib method that writes a note on a chart with an arrow pointing at one value: xy is the point, xytext where the words sit.
apply()
A pandas method that runs a function, or a lambda, on every value of a column, one at a time.
argparse
Python's standard module that reads a script's command-line arguments into positional arguments, options and flags, checks them and writes a --help page.
assert
Checks that something is true, and stops with an error if not.
astype()
A pandas method that turns a column into another type, such as text into whole numbers, and stops if any value can't be converted.
Attribute
A piece of data that belongs to an object, reached with a dot.
autoreload
A notebook kernel's tool that reloads a changed module before each cell runs, turned on with %load_ext autoreload and %autoreload 2.
Axes
In matplotlib, one chart on a Figure: the area with the data and the two axes it's measured against.
bar_label()
The matplotlib Axes method that writes each bar's value at its end, given the bars ax.bar or ax.barh handed back.
between()
A pandas method that answers True for each value from a low number to a high one, both ends included, and False for a value outside it or a missing one.
Boolean Mask
A column of True and False answers, one per row, made by a comparison; df[mask] keeps the rows where it's True.
Built-in Exceptions
The error types Python already has, like ValueError and KeyError.
capsys
A fixture pytest has built in that catches what code prints while a test runs; capsys.readouterr().out is the printed text.
Chained Comparison
Several comparisons in one go, like 18 <= age < 65.
Class
A blueprint for making objects that bundle data with actions.
Class Attribute
Data shared by every object made from a class.
Classification Report
scikit-learn's printed table of precision, recall, F1 score and support for every class, with accuracy and averages below.
Colorblind-Friendly Palette
A set of colors chosen to stay different for most color-blind readers, such as matplotlib's tableau-colorblind10 style.
concat()
pd.concat stacks tables that share their columns, one under the other; ignore_index=True numbers the stacked rows from 0 again.
Conditional Expression
Picks one of two values in one line: this if condition else that.
conftest.py
A file pytest reads before it runs the tests in its folder, for set-up every test there shares, such as choosing matplotlib's Agg backend.
Correlation
How closely two columns move together, as a number from -1 to 1; near 0 means no straight-line pattern. It doesn't show that one causes the other.
crosstab()
pd.crosstab counts how often each combination of two columns' values happens; normalize="index" turns each row into shares.
csv Module
Python's built-in module for reading and writing CSV files: csv.reader and csv.DictReader read the rows, csv.writer and csv.DictWriter write them.
Custom Exception
Your own error type, named for exactly what went wrong.
cut()
pd.cut sorts numbers into labeled bands, such as budget, mid and premium, from a list of edges.
DataFrame
pandas' table: named columns, and a label for each row called its index, like one sheet of a spreadsheet inside a program.
DataFrame.plot()
The pandas method that draws a DataFrame or Series as a chart with matplotlib (kind= picks which) and hands back the Axes it drew on.
Date Format Code
A code such as %m (month), %d (day) or %Y (four-digit year) that describes one piece of how a date is written, as in "%m/%d/%Y".
datetime Module
A standard module for working with dates and times.
Debugger (pdb)
A tool that pauses your program so you can inspect it line by line.
Deep Copy
A copy of a collection and everything inside it, fully separate.
describe()
A pandas method that sums up each number column: count, mean, standard deviation, min, the quartiles and max.
Dictionary Comprehension
Builds a new dictionary in a single line.
drop_duplicates()
A pandas method that hands back the table without its repeated rows; subset names the columns to compare and keep which copy stays.
dropna()
A pandas method that hands back the table without the rows that have missing values, or only those missing the columns named in subset.
dt accessor
Read "dot d t": the date tools of a pandas date column, such as year, month, day_name() and hour, each working on every row at once.
dtype
Short for data type: the kind of value every row of a pandas column holds, such as int64, float64 or str.
DummyClassifier
A scikit-learn model that ignores the features and guesses by a simple rule, the most common class by default, to give a baseline.
DummyRegressor
A scikit-learn model that ignores the features and always guesses one number, the training average by default, to give a baseline.
Dunder Method
A method with double underscores, like __len__, that Python calls for you.
duplicated()
A pandas method that answers True for each row that repeats a row above it, on every column or on a subset.
Encapsulation
Keeping an object's data behind its methods so it's changed the safe way.
Environment Variable
A setting stored outside your code, the safe place for passwords and keys.
Estimator
scikit-learn's word for any object that learns from data with a fit method, such as LinearRegression or StandardScaler.
Exception
Python's way of reporting an error while a program runs.
Exception Chaining (raise … from)
Raising a clearer error while keeping the original attached (raise ... from).
Exploratory Data Analysis
Looking around a new dataset to learn what's in it before answering a question: its size, columns, gaps and odd values. Often shortened to EDA.
Fairlearn
A Python package for checking and improving a model's fairness, with MetricFrame for scores by group and gap measures such as demographic parity.
Figure
In matplotlib, the whole image that gets saved, like a sheet of paper with one or more charts drawn on it.
fillna()
A pandas method that puts a stand-in, such as "Unknown" or a median, into every missing value.
filter()
Keeps only the items that pass a test.
finally
Runs after try/except no matter what, for clean-up.
fit_predict()
The scikit-learn method that fits a model and hands back its answers for the same rows in one call, as KMeans uses it.
glob()
A pathlib method that finds the files in a folder whose names match a pattern, where * stands for any text, such as week-*.csv.
groupby()
The pandas method that sorts a table's rows into groups by a column's value, so each group can be totaled or averaged on its own.
Higher-Order Function
A function that takes or returns another function.
Identity vs Equality (is vs ==)
== checks equal values; is checks the very same object.
iloc
Read "i loc": pandas' way to pick rows and columns by position, counting from 0; a slice stops before its end.
Imputation
Filling in missing values with stand-ins worked out from the data, such as a median; the statistics-class name for it.
Index
The labels of a pandas Series or DataFrame, one for each value or row, used to find them; without labels of your own, they count from 0.
Inheritance
A new class starts with everything from another class, then adds or changes things.
Instance
One object made from a class, with its own data.
ipykernel
The package that lets VS Code run a notebook's cells with a virtual environment's Python; it goes in each project's requirements.txt.
isin()
A pandas method that answers True for each value found in a list, such as the categories boots or jackets.
Iterable
Anything you can loop over with for.
joblib
A Python package whose dump() and load() save a trained model to a file and read it back; scikit-learn's usual way to save models.
json Module
Turns Python data into JSON text and back.
KMeans
scikit-learn's k-means: KMeans(n_clusters=3) learns the centers with fit(), and labels_ says which cluster each row is in.
Lambda Function
A small, nameless one-line function.
List Comprehension
Builds a new list from another collection in a single line.
Literal Type
A type hint from Python's typing module that lists the only values allowed, such as Literal["dog", "cat", "rabbit"], matched exactly.
loc
pandas' way to pick rows and columns by their labels, as df.loc[rows, columns]; a label slice includes its end.
logging Module
A standard module that records what your program does, with levels.
Magic Command
A line that starts with %, such as %autoreload 2: an instruction to a notebook's kernel, not Python, so it works only in a notebook.
Main Guard (if __name__ == "__main__")
Runs code only when the file is run directly, not when imported.
make_pipeline()
scikit-learn's shortcut that builds a Pipeline from the steps alone, naming each one after its class, such as standardscaler.
map()
Runs a function on every item in a collection.
matplotlib
The Python package most charts are drawn with: a Figure holds one or more Axes, and each Axes is one chart.
merge()
A pandas method that joins two tables side by side by a key column they share, the job a spreadsheet's VLOOKUP does; how= says which rows that don't match are kept.
Method
A function that belongs to a class, called on an object with a dot.
Method Chaining
Writing several pandas steps as one expression, one method per line, each working on the table the step before it returned.
Method Overriding
A child class writing its own version of an inherited method.
Namespace
A set of names kept separate so they do not clash.
NaT
Read "not a time": pandas' mark for a missing date or length of time.
nbconvert
The Jupyter tool that converts a notebook to other formats: jupyter nbconvert --to html story.ipynb writes a web page anyone can open.
nlargest()
A pandas method that gives the rows with the largest values in a column, such as the three most-reviewed products.
np.where()
A NumPy function that gives each row one of two values: the first where a condition is True, the second where it's False.
NumPy
The Python library for fast math on arrays of numbers; pandas keeps every column's values in a NumPy array.
NumPy Array
NumPy's container (ndarray): many values of one type, packed side by side, that does math on all of them at once.
OneHotEncoder
scikit-learn's tool for one-hot encoding: it learns each column's categories with fit() and makes a 0-or-1 column for each.
openpyxl
The Python library that reads and writes Excel .xlsx files; pandas uses it behind read_excel, so it must be installed.
OrdinalEncoder
scikit-learn's tool that turns each category into a number, 0, 1, 2 and so on, in an order you can set, such as small, medium, large.
Package
A folder of related modules you can import from.
pandas
The Python package for tables of data: it reads a CSV file into a DataFrame, then sorts, filters, totals and charts it in a line or two.
pathlib
Python's standard module for file and folder paths: a Path holds an address in parts, joins them with / and reads a file's name, stem, suffix and parent.
Pattern Matching (match/case)
match/case: compares a value against patterns and runs the first that fits.
PEP 8
Python's official style guide for how code should look.
pickle
Python's built-in module that turns objects, a trained model included, into bytes and back; never load a pickle you don't trust.
pip
The tool that installs Python packages from the internet.
pip-audit
A command that checks a project's packages against a public list of known security holes, and names the version that fixes each.
Pipeline
A scikit-learn object that chains preparing steps and a model into one, so fit() and predict() run every step in order.
pivot_table()
The pandas method that builds a pivot table from index, columns, values and aggfunc, with fill_value for empty squares and margins for totals.
plt.close()
Closes a matplotlib figure so pyplot stops keeping it in memory; a test calls plt.close(fig) when it has read the chart.
Polymorphism
Different objects answering the same method call in their own way.
predict_proba()
The scikit-learn method that gives, for each row, the classifier's probability for every class, in the order of classes_.
Private Attribute (_name)
An attribute starting with _ that others should leave alone.
pytest
The most popular Python testing tool: test functions with plain assert.
pytest.raises
pytest's way to make sure code raises an error: a test passes only when the code inside with pytest.raises(...) raises that error.
Quartile
One of the values that cut sorted data into quarters: 25% of the values are at or below the first, half at or below the second (the median).
raise
Triggers an exception on purpose.
raise_for_status()
A requests response method that raises an HTTPError when the status code is from 400 to 599, and does nothing otherwise.
Random Seed
The starting number for Python's random numbers: the same seed gives the same "random" results on every run, which is what tests need.
read_excel()
pd.read_excel loads a sheet of an Excel workbook into a DataFrame; sheet_name picks the sheet and skiprows steps over title rows.
read_json()
pd.read_json loads JSON into a DataFrame, a row for each record; lines=True reads JSON Lines.
Regular Expression
A pattern for finding or checking text.
rename()
A pandas method that hands back a table with new column names, given a dictionary of old name to new name in columns=.
requests
The Python package most programs use to call web APIs: requests.get sends a request and gives back a response to read.
requirements.txt
A file listing the packages a project needs.
reset_index()
A pandas method that numbers a table's rows 0, 1, 2 again, moving the old row labels, such as a groupby's groups, into a column, or dropping them with drop=True.
Ruff
A fast Python linter and formatter in one tool: ruff check finds likely mistakes such as unused imports, and ruff format lays out every file the same way.
self
The object a method is working on, passed in automatically.
Series
One column of a pandas table, with the table's row labels; its methods, such as sum() and max(), work on the whole column at once.
Series.map()
A pandas method that, given a dictionary, gives each value the one the dictionary pairs with it, and NaN for a value it doesn't name.
Series.replace()
A pandas method that, given a dictionary, changes only the values the dictionary names and leaves every other value as it was.
Set Comprehension
Builds a new set from another collection in a single line, written in curly braces with one expression.
Set Operations
Combining sets: | both, & shared, - only in the first.
set_index()
A pandas method that makes a column, such as a product code, the table's row labels, so loc can find rows by it.
set_ylim()
A matplotlib method that sets where a chart's y-axis starts and ends; set_ylim(bottom=0) starts it at zero.
Shallow Copy
A new collection that still shares the items inside it with the original.
Shared Axis
One axis scale used by charts side by side, so their bars and lines compare by eye; plt.subplots(sharey=True) makes one.
shutil
Python's standard module for copying and moving files and folders, such as shutil.copy2 and shutil.move; read "shell utilities".
SimpleImputer
scikit-learn's tool that fills missing values with one value per column learned in fit(), such as the column's median.
Split-Apply-Combine
What a grouping does in three steps: split the rows into groups, work out one number for each group, and combine the answers into one result.
sqlite3.IntegrityError
The exception Python's sqlite3 module raises when a change would break a rule of the data, such as a repeated value in a UNIQUE column.
Standard Deviation
How far values usually sit from their average; describe() shows it as std.
StandardScaler
scikit-learn's tool that standardizes features: fit() learns each column's mean and standard deviation, transform() rescales with them.
Step Into
The debugger's command (F11 in VS Code) that goes inside the function the highlighted line calls and pauses on its first line.
Step Over
The debugger's command (F10 in VS Code) that runs the highlighted line, a function call included, and pauses on the next one.
str accessor
Read "dot s t r": the string tools of a pandas text column, such as strip, lower and contains, each working on every value at once.
str.replace()
A pandas text tool that swaps one piece of text for another in every value of a column, such as taking " lbs" off weights.
str.split()
A pandas text tool that cuts each value at a separator, a space by default; expand=True puts the pieces in columns and n limits the cuts.
subplots()
plt.subplots makes a Figure and its Axes in one call: one chart, or a grid such as 2 by 2 picked as axs[row, column].
super()
Calls a method from the parent class, usually to add to it.
sys.argv
The list of words a Python program was started with: the script's name first, then each word typed after it, all as strings.
Test Fixture
Shared setup, like sample data, that tests can use.
Third-Party Library
Code written by others that you install with pip.
Tidy Data
The textbook name for a table where each row is one observation and each column is one kind of value.
Timedelta
pandas' type for a length of time, such as 60 days 16 hours or 27 minutes 15 seconds; subtracting two dates gives one.
Timestamp
pandas' type for one date and time, such as 2026-08-03 07:40; a datetime64 column holds one per row.
tmp_path
A fixture pytest has built in: a new, empty folder for one test, where it can write files without touching the project.
to_datetime()
pd.to_datetime turns text into dates; format= says how they were written, and errors="coerce" makes each date it can't read NaT.
to_excel()
A pandas method that saves a table as an Excel workbook (.xlsx), using the openpyxl package.
to_html()
A pandas method that gives back a table as HTML text, ready to put in a web page; index=False leaves out the row numbers.
to_numeric()
pd.to_numeric turns a column of text into numbers; with errors="coerce", each value it can't read becomes NaN.
to_period()
A .dt method that turns each date into the period it falls in, such as its month, so rows can be grouped by month.
to_timedelta()
pd.to_timedelta turns text such as "0:27:15" (hours:minutes:seconds) into a Timedelta, a length of time.
total_seconds()
A Timedelta method (.dt.total_seconds() on a column) that gives the length of time as a number of seconds.
try and except
try runs code that might fail; except catches the error instead of crashing.
Unit Test (unittest)
Code that checks one function gives the right answer.
Vectorized Operation
Math done on a whole array or column at once, with no loop of your own, such as prices * 0.9.
Version Pinning
Writing a package's exact version in requirements.txt, such as pytest==9.2.0, so every computer installs the same one.
Virtual Environment
A project's own private set of installed packages.
Walrus Operator
The := operator: saves a value and uses it in the same line.

Every Python term, by level