Exploratory Data Analysis Framework

You are a senior data scientist conducting an exploratory data analysis. Analyze the following dataset context and produce a comprehensive EDA framework.

Dataset Description: [DESCRIBE YOUR DATASET - e.g., customer transactions, sensor readings, survey responses]
Number of Records: [APPROXIMATE ROW COUNT]
Number of Features: [APPROXIMATE COLUMN COUNT]
Data Source: [WHERE THE DATA COMES FROM]
Business Context: [WHAT BUSINESS PROBLEM THIS DATA RELATES TO]
Target Variable (if any): [VARIABLE YOU WANT TO PREDICT OR UNDERSTAND]

Produce a structured EDA plan covering these six sections:

1. DATA PROFILING AND SUMMARY STATISTICS
Outline the steps to generate a complete profile of the dataset. Include checks for data types, missing value percentages per column, unique value counts, and basic descriptive statistics such as mean, median, standard deviation, skewness, and kurtosis for numeric features. For categorical features, detail frequency distributions and cardinality checks. Recommend specific Python functions or libraries such as pandas-profiling, sweetviz, or manual pandas describe methods. Provide code snippets for each profiling step and explain what to look for in the output.

2. DATA QUALITY AND CLEANING ASSESSMENT
Describe a systematic approach to identifying data quality issues. Cover duplicate detection strategies, outlier identification using IQR and z-score methods, inconsistent formatting detection for dates and strings, and validation of value ranges against domain knowledge. Provide a decision framework for handling missing data including when to drop, impute with mean or median, use forward fill, or apply more sophisticated imputation like KNN or MICE. Include a data quality scorecard template.

3. UNIVARIATE ANALYSIS
Detail the visualizations and statistical tests for examining each variable individually. For numeric variables, recommend histograms with KDE overlays, box plots, and QQ plots to assess normality. For categorical variables, suggest bar charts with percentage annotations and Pareto charts. Explain how to interpret each visualization and what red flags to watch for. Provide matplotlib and seaborn code templates for each chart type with proper formatting and labeling.

4. BIVARIATE AND MULTIVARIATE ANALYSIS
Outline the approach for examining relationships between variables. Cover correlation matrices with heatmaps for numeric pairs, chi-square tests for categorical associations, point-biserial correlation for mixed types, and scatter plot matrices for key variable subsets. Describe how to create and interpret pair plots, grouped box plots, and stacked bar charts. Include guidance on feature interaction detection and how to identify collinearity issues that could affect downstream modeling.

5. TEMPORAL AND SEGMENTATION ANALYSIS
If the data has time components, describe trend analysis, seasonality detection, and change point identification approaches. For non-temporal data, focus on segmentation analysis using groupby operations to compare subpopulations. Recommend rolling statistics, decomposition plots, and cohort-based comparisons. Explain how to identify meaningful segments and validate whether observed differences are statistically significant using appropriate hypothesis tests.

6. KEY FINDINGS AND NEXT STEPS
Provide a template for summarizing EDA findings into actionable insights. Include a format for documenting the top discoveries, data quality concerns that need resolution, feature engineering opportunities identified, recommended modeling approaches based on the data characteristics, and specific hypotheses to test. Create a prioritized action list for the next phase of the data science workflow.