Data exploration
Data exploration is the systematic examination and summarization of data to understand its structure, quality, distributions, relationships, patterns, and anomalies.
Machine Learning for Tabular Data: XGBoost, Deep Learning, and AI
Exploration is the investigation of previously untested regions or configurations in a search space.
- How it is used
- Guides cleaning, feature decisions, target checks, and validation preparation before tabular modeling.
- How deep it goes
- Develops a staged workflow from univariate profiles through bivariate associations and multivariate projections, without prescribing a fixed EDA blueprint.
- Examples
- Uses Auto MPG feature profiles and an Airbnb regression target with extreme-value filtering.
- What you can do after
- Apply it with pandas summaries and plots to identify data types, anomalies, missingness, redundancy, and target extremes before modeling.
Statistics Slam Dunk
Exploratory data analysis is the systematic examination of data using summaries, visualizations, and other descriptive methods before formal modeling or inference.
- How it is used
- Establishes an initial read on prepared data before statistical testing, modeling, or further analysis.
- How deep it goes
- Combines stepwise preparation, descriptive statistics, and visual inspection, then defers formal inference and modeling.
- Examples
- Uses NBA draft win shares and basketball betting totals, with profiling reports, histograms, density plots, correlations, and bar charts.
- What you can do after
- Apply it via R and DataExplorer to profile datasets, inspect structure and missingness, and generate descriptive tables and plots.
Effective Data Analysis
Exploratory data analysis is the systematic examination of a dataset to understand its distributions, patterns, anomalies, and limitations before formal modeling or testing.
- How it is used
- Guides early understanding of datasets and prepares choices about summaries, statistical tests, and models.
- How deep it goes
- Details visual, quality, and summary-statistic checks, then extends exploration into feature derivation; defers detailed instruction on statistical concepts and methods.
- Examples
- Uses daily rat-sighting and weather data, monthly boxplots, and 90-day Help Center summaries to illustrate exploratory choices.
- What you can do after
- Practice/perform it by checking distributions and data quality, exploring predictors, and documenting the choices behind summaries.
Read how Effective Data Analysis explains data exploration →
Data Without Labels
Exploratory data analysis is the analytical examination of a dataset to summarize its characteristics, identify patterns, and generate initial insights.
- How it is used
- Generates business insights and checks data behavior before modeling, with findings discussed with stakeholders.
- How deep it goes
- Details variable inspection and a short Python workflow; defers a full, commented analysis to a separate notebook.
- Examples
- Uses vehicle-data checks before k-means and customer or subscriber datasets to illustrate patterns, distributions, and variable screening.
- What you can do after
- Practice/perform initial dataset checks and visual inspection to develop insights that guide business discussion and modeling.
Optimization Algorithms: AI techniques for design, planning, and control problems
A search strategy that investigates new or insufficiently tested actions to gather information about their potential value.
- How it is used
- Uses exploration to test unfamiliar feasible solutions or actions instead of relying only on currently promising choices.
- How deep it goes
- Defines the trade-off and works through simulated selection strategies, parameters, and regret without developing a formal proof.
- Examples
- Uses unfamiliar restaurants and a simulated three-slot-machine problem with random, epsilon-greedy, and UCB strategies.
- What you can do after
- Apply it via Python and NumPy to compare bandit exploration strategies using average regret and adjustable parameters.
Delivering Machine Learning Projects: From design to deployment
Exploratory data analysis is the examination of data to characterize its structure, quality, patterns, and anomalies before or during modeling.
- How it is used
- Uses EDA before modeling to test data credibility, characterize available information, and judge whether modeling can meet project requirements.
- How deep it goes
- Develops question-led planning, specific integrity checks, and a shared report rather than a detailed statistical recipe.
- Examples
- Contrasts naïvely treated sales with deduplicated, normalized sales and checks aggregates against published revenue.
- What you can do after
- Practice/perform a ticketed EDA that checks plausibility, distributions, and integrity, then shares findings for modeling decisions.
Read how Delivering Machine Learning Projects: From design to deployment explains data exploration →
Starting Data Analytics with Generative AI and Python
Exploratory data analysis is the process of examining data with summaries, statistical methods, domain knowledge, and visualizations to discover structure, patterns, trends, and potential issues.
- How it is used
- Supports data-quality assessment by examining distributions, visualizations, variable meaning, and consistency.
- How deep it goes
- Defines four inspection steps and extends exploration to time-series outliers and cyclicity.
- Examples
- Uses order reviews to inspect missing values, score distributions, monthly review counts, and score proportions over time.
- What you can do after
- Evaluate data by checking distributions, visualizing patterns over time, and questioning what variables represent.
Read how Starting Data Analytics with Generative AI and Python explains data exploration →
Designing Deep Learning Systems: A guide for software engineers
Data exploration is the systematic examination of available data to understand its contents, quality, distributions, and useful patterns.
- How it is used
- Starts deep learning work by assembling and examining training data against business requirements.
- How deep it goes
- Works through collection, labeling, cleaning, and iterative experiment selection; defers formal data pipelines until productionization.
- Examples
- Uses a natural-language customer-support service and researcher or data-scientist workflows over available data.
- What you can do after
- Apply it via scripts, notebooks, and data-management services to prepare data and compare experimental runs.
Read how Designing Deep Learning Systems: A guide for software engineers explains data exploration →
Everyday Data Visualization
Data exploration is the process of examining data to discover patterns, relationships, or useful questions.
- How it is used
- Facilitates user investigation of visualized data through purposeful, accessible interaction.
- How deep it goes
- Explains interaction choices, user mental models, visibility, feedback, errors, devices, and accessibility, but not analytical algorithms.
- Examples
- Uses a bank website, a fitness-tracker app, filters, toggles, tooltips, and dashboard design as concrete exploration examples.
- What you can do after
- Apply interaction patterns that help users explore visualized data without hiding essential information or excluding device users.
Read how Everyday Data Visualization explains data exploration →
Financial AI in Practice
Exploratory data analysis is the examination of data to understand its structure, quality, and distributions before further analysis or modeling.
- How it is used
- Inspects financial datasets before modeling or transforming transaction tables into graphs.
- How deep it goes
- Demonstrates summary checks, cleaning context, and distribution plots, but stops short of deep feature selection.
- Examples
- Uses credit-default proportions, categorical countplots, target-colored numeric histograms, and transaction-table summaries.
- What you can do after
- Evaluate/understand datasets by spotting missingness, imbalance, rare categories, outliers, and distribution differences.
Read how Financial AI in Practice explains data exploration →
The AI Pocket Book
In reinforcement learning, the practice of trying random actions to gather information.
- How it is used
- Names the information-gathering phase of reinforcement learning before exploitation takes over.
- How deep it goes
- Defines exploration, illustrates it with ads, and explains its tradeoff with exploitation without going into algorithms.
- Examples
- Uses random ad serving and click tracking to show how exploration learns user preferences.
- What you can do after
- Evaluate when random actions are still needed to discover new preferences.
Graph Neural Networks in Action
The process of examining and summarizing data to discover patterns, structure, relationships, and potential anomalies.
- How it is used
- Guides dataset preparation and topology checks before selecting and training graph neural network models.
- How deep it goes
- Moves from loading features and adjacency data to implementing graph metrics, but does not detail downstream model construction.
- Examples
- Uses the YelpChi dataset, MATLAB loading code, adjacency lists, NetworkX graphs, and tabulated topology measurements.
- What you can do after
- Apply Python, SciPy, and NetworkX workflows to load graph data and diagnose its structural properties before training.
Read how Graph Neural Networks in Action explains data exploration →
Statistics Every Programmer Needs
Exploratory analysis is the investigation of data to discover patterns, relationships, anomalies, and plausible hypotheses.
- How it is used
- Uses exploratory analysis to assess data quality and guide appropriate statistical methods before modeling.
- How deep it goes
- Defines the technique and specifies checks for outliers, normality, and relationships, without detailing their individual tests.
- Examples
- Demonstrates regression preparation through outlier detection, normality testing, relationship analysis, and Benford distribution comparison.
- What you can do after
- Diagnose data quality and potential assumption problems before fitting statistical models or interpreting analytical results.
Read how Statistics Every Programmer Needs explains data exploration →
Related concepts
- datasetuses · A dataset is an organized collection of related data items, records, or observations treated as a unit for storage, processing, analysis, or computation.
- visualizationuses · Visualization is the representation of data, information, or concepts in visual form to support inspection, interpretation, analysis, or communication.
- pandasused by · pandas is an open-source Python library that provides labeled data structures and tools for manipulating and analyzing tabular and other structured data.
- exploitationcontrasts with · In optimization and decision-making, exploitation is selecting an option currently believed, based on available information, to yield high value.
- DataExplorerused by · DataExplorer is an R package that automates exploratory data analysis by generating data profiling summaries and visualizations.
- unstructured datauses · Unstructured data is information that lacks a predefined schema or consistently organized fields.
These guides are generated from the text of Manning books by Manning's knowledge graph and link to the sections they draw on. The full books are available on liveBook and at manning.com.