aNet.ai
Data Science/Glossary
0day streak0XP · Lv 1
DSData Science0/53 lessonsMLMachine Learning0/151 lessons

Glossary

135 data science terms in plain language, each linked to the topic that teaches it.

135 terms

A

Accuracy
Values match the true source.
3. Data Cleaning and Quality
Approximation
A useful estimate, not an exact promise.
2. Describing Data
argmin
Position of the smallest value.
4. From Clean Data to Useful Predictions
Association
Two measurements vary together.
1. Thinking Like a Data Scientist
Audit log
A record of changes and reasons.
3. Data Cleaning and Quality
Availability time
When the system could access its recorded information.
4. From Clean Data to Useful Predictions

B

Bar chart
Category counts or rates.
5. Exploratory Data Analysis
Baseline
Measured before an action.
5. Exploratory Data Analysis
Bimodal
Two main peaks.
2. Describing Data
Bin
An interval used by a histogram.
2. Describing Data
Binary outcome
A value such as 1 for pass and 0 for fail.
1. Thinking Like a Data Scientist
Bound
An extreme possible value given the available information.
3. Data Cleaning and Quality
Bounds
Extremes obtained by assigning all unknown outcomes one way or the other.
5. Exploratory Data Analysis

C

Case code
Identifies your generated version.
2. Describing Data
Causal effect
Changing one thing changes an outcome.
1. Thinking Like a Data Scientist
Coercion
Replacing an unparseable value with a missing marker.
3. Data Cleaning and Quality
Completeness
Needed data are present.
3. Data Cleaning and Quality
Confounder
A common cause that can affect both.
1. Thinking Like a Data Scientist
Coordinate
One position in a vector.
4. From Clean Data to Useful Predictions
Copy
An extra record of the same thing.
3. Data Cleaning and Quality
Correction
A verified replacement.
3. Data Cleaning and Quality
Cutoff
The latest information the system may use.
4. From Clean Data to Useful Predictions

D

Data
Recorded information.
1. Thinking Like a Data Scientist
Data dictionary
Definitions and rules for each field.
1. Thinking Like a Data Scientist
Denominator
The count you divide by.
1. Thinking Like a Data Scientist
Deviation
Value minus mean.
2. Describing Data

E

EDA
Checking and exploring data before making strong claims.
5. Exploratory Data Analysis
Error
Wrong/evaluated.
4. From Clean Data to Useful Predictions
Euclidean distance
Square root of the sum of squared coordinate differences.
4. From Clean Data to Useful Predictions
Event time
When something happened.
4. From Clean Data to Useful Predictions
Exclude for analysis
Leave out a measurement for a clearly limited calculation.
3. Data Cleaning and Quality

F

Feature
An input used to predict.
4. From Clean Data to Useful Predictions
Feature engineering
Creating inputs from recorded information.
4. From Clean Data to Useful Predictions
Feature order
The meaning assigned to each position.
4. From Clean Data to Useful Predictions
Fence
Calculated review limit.
2. Describing Data
Fit
Learn parameters from training data.
4. From Clean Data to Useful Predictions
Future leakage
Later information in early prediction.
4. From Clean Data to Useful Predictions

H

Histogram
Numeric values grouped into intervals.
5. Exploratory Data Analysis

I

Imputation
Filling missing values using a stated rule.
4. From Clean Data to Useful Predictions
Impute
Fill using an assumption or model.
3. Data Cleaning and Quality
Interpretation
Explaining what a number means for the question.
2. Describing Data
Invalid value
Breaks a documented rule.
3. Data Cleaning and Quality
IQR
Q3−Q1.
2. Describing Data

K

Key
The value or set of values that identifies one row.
1. Thinking Like a Data Scientist
Known-outcome rate
Successes among recorded outcomes.
5. Exploratory Data Analysis

L

Left join
Keep all left-side records and add matches.
1. Thinking Like a Data Scientist
Linear
A straight-line pattern.
5. Exploratory Data Analysis

M

Many-to-one
Many left records can match one right record.
1. Thinking Like a Data Scientist
MAR
After using known information, missingness no longer depends on the missing value.
3. Data Cleaning and Quality
MCAR
Missingness unrelated to the values.
3. Data Cleaning and Quality
Mean
Total divided by observed count.
2. Describing Data
Measurement validity
Whether it measures the intended idea.
1. Thinking Like a Data Scientist
Median
Middle after sorting.
2. Describing Data
Min–max rule
(value−training minimum)/(training maximum−training minimum).
4. From Clean Data to Useful Predictions
Missing
Amount unknown.
3. Data Cleaning and Quality
Missing count
The eligible measurements not available.
2. Describing Data
MNAR
That dependence remains.
3. Data Cleaning and Quality
Mode
Most frequent value.
2. Describing Data
Model
A rule that finds patterns or makes predictions.
1. Thinking Like a Data Scientist
Monotone
Preserves the order of values.
4. From Clean Data to Useful Predictions

N

Nearest neighbour
The training record with the smallest distance under the saved representation.
4. From Clean Data to Useful Predictions
Neighbour
A compared training record.
4. From Clean Data to Useful Predictions
New
Receives a prediction without a known outcome.
4. From Clean Data to Useful Predictions
Nominal
Categories with no order.
1. Thinking Like a Data Scientist
Normal model
A particular bell-shaped distribution.
2. Describing Data
Not applicable
The measurement does not belong to this person or situation.
3. Data Cleaning and Quality
Numeric
A measured or counted amount.
1. Thinking Like a Data Scientist

O

One-hot
A vector of category indicators.
4. From Clean Data to Useful Predictions
One-hot positions
CS, Economics and Psychology.
4. From Clean Data to Useful Predictions
Ordinal
Ordered categories without guaranteed equal gaps.
1. Thinking Like a Data Scientist
Outlier flag
A value worth checking.
3. Data Cleaning and Quality
Overfitting
Learning details that do not carry well to new data.
4. From Clean Data to Useful Predictions

P

Parsing
Reading text into a structured value.
3. Data Cleaning and Quality
Pattern
A visible feature that still needs interpretation.
5. Exploratory Data Analysis
Pearson r
A unitless measure between −1 and 1 when both variables vary.
5. Exploratory Data Analysis
Percentage-point difference
Subtract two percentages.
5. Exploratory Data Analysis
Percentile
Position in an ordered distribution.
2. Describing Data
Pipeline
Ordered processing steps.
4. From Clean Data to Useful Predictions
Population
The people or things you want to describe.
1. Thinking Like a Data Scientist
Population-style SD
The specified scaler convention using n.
4. From Clean Data to Useful Predictions
Post-intervention
Measured after an action and possibly affected by it.
5. Exploratory Data Analysis
Provenance
Where data came from and what happened to it.
1. Thinking Like a Data Scientist
Provenance flag
A marker showing where a value came from.
4. From Clean Data to Useful Predictions
Proxy
A stand-in measurement.
1. Thinking Like a Data Scientist

Q

Q1
25th percentile.
2. Describing Data
Q2
Median.
2. Describing Data
Q3
75th percentile.
2. Describing Data

R

Range
Maximum−minimum.
2. Describing Data
Rate
Successes divided by eligible cases.
5. Exploratory Data Analysis
Recover
Obtain the verified observation.
3. Data Cleaning and Quality
Reference group
The group that supplies the mean and SD.
2. Describing Data
Registration
One student's course enrolment.
5. Exploratory Data Analysis
Relative percentage change
Divide the difference by the starting percentage.
5. Exploratory Data Analysis
Reliability
Consistent measurement.
1. Thinking Like a Data Scientist
Repeated observation
A new week, course or event.
3. Data Cleaning and Quality
Reply share
That group's replies divided by all replies.
3. Data Cleaning and Quality
Response rate
Replies divided by that group's invitations.
3. Data Cleaning and Quality
Roster
The list of expected people or records.
3. Data Cleaning and Quality
Rounding
Shortening a number for display.
2. Describing Data
Row removed
The entire record leaves the working table.
3. Data Cleaning and Quality
Row unit, or grain
What one row represents.
1. Thinking Like a Data Scientist

S

Sample
The people actually observed.
1. Thinking Like a Data Scientist
Sampling frame
The list used to reach people.
1. Thinking Like a Data Scientist
Saved parameters
The training values used to transform later records.
4. From Clean Data to Useful Predictions
Scatter plot
Pairs of numeric values.
5. Exploratory Data Analysis
SD
Square root of variance.
2. Describing Data
Sensitivity scenario
A stated assumption to explore.
3. Data Cleaning and Quality
Simpson's paradox
Reversal between pooled and within-group comparisons.
5. Exploratory Data Analysis
size
All group rows. count: nonmissing outcomes. sum: recorded successes for a 0/1 outcome.
5. Exploratory Data Analysis
Skew
A longer tail in one direction.
2. Describing Data
Small cell
A subgroup with few recorded cases.
5. Exploratory Data Analysis
Standardise
Subtract mean, divide by SD.
4. From Clean Data to Useful Predictions
Stratify
Compare within groups.
5. Exploratory Data Analysis

T

Target
What you want to predict.
4. From Clean Data to Useful Predictions
Target leakage
The answer enters inputs.
4. From Clean Data to Useful Predictions
Target population
Everyone the conclusion concerns.
1. Thinking Like a Data Scientist
Test
Checks predictions against known outcomes.
4. From Clean Data to Useful Predictions
Test leakage
Held-out data influence fitting or choices.
4. From Clean Data to Useful Predictions
Threshold tree
Makes splits such as “value ≤ a cutoff”.
4. From Clean Data to Useful Predictions
Time window
The exact period included in a calculation.
4. From Clean Data to Useful Predictions
Training
Learns the procedure.
4. From Clean Data to Useful Predictions
Training median
The middle observed training value.
4. From Clean Data to Useful Predictions
Transform
Apply the saved parameters.
4. From Clean Data to Useful Predictions

U

Unknown category
A value not in the fitted category list.
4. From Clean Data to Useful Predictions

V

Valid n
The number of observed values used.
2. Describing Data
Validation
Data used to choose settings.
4. From Clean Data to Useful Predictions
Validity
Values follow rules.
3. Data Cleaning and Quality
Value flagged
One field is marked while the rest of the row can stay.
3. Data Cleaning and Quality
Variable
A recorded feature, such as score or logins.
1. Thinking Like a Data Scientist
Variance
Average squared deviation under the chosen divisor.
2. Describing Data

W

Weight
The share of cases in a subgroup.
5. Exploratory Data Analysis
Weighted mean
An average that uses the relevant group sizes.
5. Exploratory Data Analysis
Whisker
Most extreme observed value still inside a fence.
2. Describing Data
Working
The intermediate calculations that justify your result.
2. Describing Data

Z

Zero
Measured absence of an amount.
3. Data Cleaning and Quality
aNet.ai · Data Science Fundamentals · Weeks 1–5 · practice material, not an official examProgress saved in this browser