Data 101 Week 1

Structural concepts, data types, summary metrics, and data quality issues

Column

A vertical strip in a table representing one attribute

Attribute

What a column is called when a row is treated as an object — a property attached to that record

Feature

A column used as input to a predictive model

Variable

A column in a statistical model

Field

The database-world term for a column

Value

The smallest unit of data in a table — a single cell's content

Cell

The spreadsheet term for a value

Observation

A statistician's term for an entire row — everything measured about one subject

Data point

A value plotted on a chart

Tuple

The formal database-theory term for a row — an ordered sequence of values

Record

The everyday database/software-engineering term for a row

Instance

The machine-learning term for a row

Dataset

The umbrella term for the entire table — every row and column together

Data frame

A rectangular dataset with labeled columns

Schema

The blueprint listing every column's name and type

Cardinality

The count of distinct values in a column

Distribution

The overall shape formed by all values in a column

Missingness

How much of a column or dataset is empty or null

Granularity

The level of detail each row represents (e.g. daily vs. hourly)

Sparsity

How much of a dataset is empty

Dimensionality

The number of columns/features a dataset has

Integer

A whole-number numeric type (e.g. age)

Double/Float

A decimal-precision numeric type (e.g. GPA)

String

A text type that can't meaningfully be averaged but can be alphabetized or searched

Factor

A categorical type with a fixed set of ordered levels (e.g. class year)

Boolean

A type holding only true/false (or 1/0) values

NA/NULL

A recognized marker for a missing value

Mean

The sum of all values divided by count — highly sensitive to outliers

Median

The middle value when sorted — robust against outliers

Mode

The most frequent value — the only central-tendency measure for categorical data

Variance

The average of squared deviations from the mean

Standard deviation

The square root of variance

Range

Maximum minus minimum — relies only on the two extremes

Interquartile Range (IQR)

Q3 minus Q1 — the spread of the middle 50%

Skewness

The directional asymmetry of a distribution

Kurtosis

Tail weight and peakedness of a distribution relative to normal

Outlier

A value sitting unusually far from the rest of the dataset

Population

The entire group under study

Sample

A subset of the population actually observed

Sample size (n)

The number of observations in a sample

Projection

The formal term for selecting specific columns

Split-apply-combine

The pattern behind group-by operations: split into groups

Misspelling

A text-pollution issue where variant spellings split true counts during aggregation

Case inconsistency

When identical values appear with different capitalization across rows

Extra spaces

Invisible leading/trailing/internal whitespace that breaks exact-match filters

Encoding issue

When text is read with the wrong character encoding

Placeholder noise

Missing data disguised as valid text like "unknown" or "N/A"

Exact duplicates

Identical rows appearing multiple times byte-for-byte

Near duplicates

Rows referring to the same entity with slightly different text (e.g. "Jon" vs "John")

Redundant columns

Separate columns storing identical information under different names

Label variation

A single category recorded under multiple text representations (e.g. "USA" vs "US")

Contradictory fields

Two columns in the same row holding mutually exclusive value

plot()

Creates base R scatterplots and general plots

boxplot()

Creates box-and-whisker plots showing distributions across groups using formula syntax

barplot()

Generates bar graphs from a vector of height values

mosaicplot()

Creates a mosaic plot to display categorical data relationships

ggplot()

Initializes a ggplot2 visualization object with input data

geom_bar()

Adds a bar chart layer to a ggplot object

geom_boxplot()

Adds a boxplot layer to a ggplot object

aes()

Constructs aesthetic mappings connecting data columns to visual properties (x