PCED key terms

The complete set of all units or records you want to draw conclusions about, such as every customer in a database or every sensor reading in a period.

A subset of a population that is actually measured or analyzed, used when collecting data from every unit is impractical or impossible.

A systematic distortion that occurs when the units selected for a sample differ in relevant ways from the population, so results do not generalize.

The degree to which a sample mirrors the composition and variation of the population on characteristics that matter for the question being studied.

Data organized into a defined schema of rows and columns or key-value pairs, such as a CSV table or a database table, that is easy to query programmatically.

Data without a predefined tabular model, such as free text, images, audio, or log messages, which usually needs parsing before it can be analyzed.

The category that determines which values a field can hold and which operations are valid, for example integer, float, string, boolean, or date.

A named reference in a program that points to a value stored in memory, allowing that value to be reused, reassigned, and passed to functions.

An ordered, mutable Python sequence that holds items by position and supports indexing, slicing, appending, and iteration over its elements.

A mutable Python mapping of unique keys to values that allows fast lookup by key rather than by numeric position, useful for counting and grouping.

An ordered, immutable Python sequence often used for fixed records such as coordinate pairs or rows that should not be modified after creation.

An unordered Python collection of unique hashable elements that supports fast membership tests and mathematical operations such as union and intersection.

A compact Python expression that builds a new list by applying an expression to each item of an iterable, optionally filtering with a condition.

A lazy Python expression that yields items one at a time from an iterable, saving memory when processing large data streams that need not be stored.

The order in which statements execute, directed by conditional branches such as if and elif and by loops such as for and while.

A construct that repeats a block of statements, either over items of an iterable with for or while a condition remains true.

A named, reusable block of code that accepts parameters, performs a task, and optionally returns a value, improving clarity and avoiding duplication.

The use of try, except, else, and finally blocks to detect runtime errors and respond to them without crashing the whole program.

Opening, reading, writing, and closing files in Python, typically with a with statement so the file is closed automatically even if an error occurs.

A plain-text format that stores tabular data as rows of values separated by a delimiter, usually a comma, with an optional header row.

A Python standard-library module providing reader and writer objects, plus DictReader and DictWriter, for parsing and producing delimited text files.

The character or string that separates fields within a record in a delimited file, commonly a comma, tab, semicolon, or pipe.

An absent or unrecorded entry in a dataset, represented in Python by None or NaN, which must be handled before aggregation or modeling.

An observation that lies far from the rest of the distribution, which may reflect genuine variation, a measurement error, or a data-entry mistake.

A row that repeats the same entity or measurement already present in the dataset, often inflating counts and skewing summary statistics if not removed.

The process of detecting and correcting inaccurate, incomplete, duplicated, or inconsistently formatted records so analysis results are trustworthy.

Converting a value from one data type to another, such as turning a numeric string into an integer, which may fail or silently alter meaning.

Filling in missing values with a substitute such as the mean, median, mode, or a predicted value, rather than discarding the affected records.

Summary measures that characterize a dataset, including counts, central tendency such as mean and median, and spread such as standard deviation and range.

The arithmetic average of a set of values, computed by summing them and dividing by the count, and sensitive to extreme values.

The middle value of an ordered dataset, or the average of the two middle values, which is more robust to outliers than the mean.

A measure of how far values typically lie from the mean, equal to the square root of the variance and expressed in the original units.

A Python standard-library module offering functions such as mean, median, mode, stdev, variance, and correlation for small to moderate datasets.

A fixed-type, multidimensional container from the NumPy library that supports fast element-wise arithmetic, slicing, and vectorized operations on numeric data.

Performing an operation on an entire array at once instead of looping element by element, which produces shorter code and much faster execution.

NumPy's rule for combining arrays of different shapes by conceptually stretching the smaller one, such as subtracting a per-column mean from every row.

A collections class that counts occurrences of elements and returns them as a dictionary-like mapping of item to frequency, with methods for most common items.

A collections subclass of dict that supplies a default value for missing keys, commonly used to group records into lists without pre-initializing each key.

A Python standard-library module providing date, time, timedelta, and datetime objects for parsing, formatting, comparing, and arithmetic on dates and times.

Converting a textual date or time into a datetime object, typically with strptime and an explicit format string, so it can be sorted and compared.

A regional offset from UTC that determines local clock time, which must be handled explicitly to avoid misaligned or double-counted time-based analysis.

Splitting records into groups by one or more key fields and computing a summary such as count, sum, or average for each group.

Reshaping, filtering, deriving, or normalizing data so it fits the analysis, for example adding computed columns or converting wide layouts to long ones.

Encoding data values as visual properties such as position, length, or color so patterns, comparisons, and anomalies become easier to perceive.

A chart that represents categorical values as bars whose lengths encode magnitude, well suited to comparing a modest number of discrete categories.

A chart that bins a numeric variable into intervals and shows the count or frequency per bin, revealing the shape and spread of a distribution.

The proportion of a graphic's ink devoted to actual data rather than decoration, gridlines, or redundant labels; maximizing it reduces visual clutter.

Decorative or distracting elements in a chart, such as heavy gridlines, 3D effects, or gratuitous images, that add no information and can mislead.

An axis whose scale, starting point, or breaks distort perceived differences, such as a bar chart baseline that does not begin at zero.

Structuring an analysis into a clear narrative with context, a central finding, supporting evidence, and a recommended action for a specific audience.

The people who will receive an analysis, whose technical background and decisions determine the appropriate level of detail, vocabulary, and chart complexity.

A short opening section that states the main finding and its implications up front, letting busy readers grasp the conclusion before the supporting detail.

Permission obtained from people before their data is collected or reused, given with a clear understanding of purpose, scope, and possible consequences.

Removing or transforming identifiers so individuals cannot reasonably be re-identified, noting that combining quasi-identifiers can still enable re-identification.

The principle of collecting and retaining only the data actually needed for a stated purpose, reducing exposure and privacy risk over time.

Distortion caused by how records enter the dataset, for example analyzing only customers who completed a survey or only rows that passed a filter.