fbla

Data Science

Using data, statistics, and computing to gain insights and make decisions.

Data

Raw facts and numbers.

Information

Processed data with meaning.

Algorithm

Step-by-step procedure to solve a problem.

Model

A mathematical representation of a real-world process.

Define Problem

First step in data science: decide what question you're answering.

Collect Data

Gather data from sources.

Clean Data

Fix missing values, errors, and duplicates.

Explore Data

Use charts and stats to look for patterns.

Build Model

Fit algorithms to data.

Evaluate Model

Measure accuracy, precision, recall, etc.

Communicate Results

Share insights clearly.

Qualitative Data

Descriptive categories (colors, names).

Quantitative Data

Numerical values.

Discrete Data

Countable integers.

Continuous Data

Real-number values.

Nominal Data

Categories with no order (eye color).

Ordinal Data

Ordered categories (rankings).

Interval Data

Numeric, no true zero (°C).

Ratio Data

Numeric, true zero (height, money).

Mean

Average.

Median

Middle value.

Mode

Most frequent value.

Range

Max minus min.

Variance

Measure of spread.

Standard Deviation

Spread of data; square root of variance.

Correlation

Strength/direction of relationship (−1 to +1).

Correlation does NOT imply causation

Relationship does not mean one causes the other.

Bar Chart

Compares categories visually.

Histogram

Shows distribution of numerical data.

Scatter Plot

Shows relationship between two numeric variables.

Line Chart

Shows trends over time.

Box Plot

Shows quartiles and outliers.

Database

Organized collection of data.

Table

Rows and columns storing data.

Primary Key

Unique identifier for each row.

Foreign Key

Field linking one table to another.

SELECT

SQL command to choose columns.

FROM

SQL command to choose table.

WHERE

SQL command to filter rows.

ORDER BY

SQL command to sort rows.

JOIN

SQL command to combine tables.

Artificial Intelligence (AI)

Machines performing tasks requiring human intelligence.

Machine Learning (ML)

Algorithms that learn patterns from data.

Deep Learning

ML using neural networks with many layers.

Supervised Learning

Labeled data; classification and regression.

Unsupervised Learning

Unlabeled data; clustering or dimensionality reduction.

Reinforcement Learning

Learning through rewards and punishments.

Linear Regression

Predicts numeric values.

Logistic Regression

Classification algorithm (yes/no).

Decision Tree

Tree-based decisions.

Random Forest

Many decision trees combined.

SVM (Support Vector Machine)

Finds best separating boundary.

k-NN (k Nearest Neighbors)

Predicts from closest examples.

k-Means Clustering

Unsupervised clustering algorithm.

PCA (Principal Component Analysis)

Reduces number of features.

Neural Network

Model with layers of connected neurons.

Input Layer

Takes in features.

Hidden Layer

Learns patterns.

Output Layer

Produces final prediction.

Weights

Values the model learns.

Backpropagation

Method neural nets use to adjust weights.

Accuracy

Correct predictions / total predictions.

Precision

TP / (TP + FP).

Recall

TP / (TP + FN).

Confusion Matrix

Table of true/false positives/negatives.

MAE (Mean Absolute Error)

Average absolute difference from true values.

MSE (Mean Squared Error)

Measures error squared.

R² Score

How well a regression fits the data.

Remove Duplicates

Delete repeated rows.

Handle Missing Values

Fill with mean/median or drop rows.

Normalize

Scale values to similar ranges.

Standardize

Convert data to mean 0, SD 1.

One-Hot Encoding

Turn categories into 0/1 columns.

Outlier Detection

Identify extreme values.

AI Bias

Unfair outcomes for certain groups.

AI Transparency

Ability to explain model decisions.

Data Privacy

Protecting user data.

AI Accountability

Responsibility for system actions.

Fairness in AI

Equal treatment regardless of group.

Overfitting

Model memorizes training data, performs poorly on new data.

Underfitting

Model too simple; misses patterns.

Training Set

Data model learns from.

Test Set

Data used to evaluate performance.

Feature

Input variable.

Label

Target variable you're predicting.