basicdatainfo 0.1.0

One-line automated EDA for pandas DataFrames: column analysis, cleaning, visualization, and HTML reports.

basicdatainfo

One-line automated EDA for pandas DataFrames.

Send it a DataFrame and it will:

• Detect each column's type (numeric, categorical, datetime, boolean, text)
• Compute summary statistics for every column
• Optionally clean the data (clean=True): drop duplicate rows, strip whitespace, auto-parse date columns, fill missing values, optionally cap outliers
• Automatically engineer date features (feature_engineer=True, default): if any column is (or becomes) a datetime column, it automatically derives year, quarter, month, month_name, week_of_year, day_of_month, day_of_week, day_name, is_weekend, is_month_start/end, is_quarter_start/end, and hour (if a time component is present) — and analyzes/visualizes all of these new columns too
• Automatically generate visualizations: histograms + boxplots for numeric columns, an extra "count by value" bar chart for low-cardinality numeric columns (e.g. the engineered month/day-of-week/quarter features), a "records over time" trend chart for datetime columns, bar charts for categorical columns, a missing-values chart, and a correlation heatmap
• Either print a summary + show the plots (default), or generate a single self-contained HTML report (html=True)

Install

pip install -e .

(or build a wheel with python -m build and pip install dist/*.whl)

Usage

import pandas as pd
import basicdatainfo

df = pd.read_csv("data.csv")

# 1. Just analyze + show plots inline (e.g. in Jupyter)
basicdatainfo.analyze(df)

# 2. Auto-clean the data first, then analyze
result = basicdatainfo.analyze(df, clean=True)
clean_df = result["df"]

# 3. Generate a standalone HTML report instead of inline output
basicdatainfo.analyze(df, clean=True, html=True, output_path="report.html")

analyze() parameters

Parameter	Default	Description
clean	False	Auto-clean the dataframe before analysis
feature_engineer	True	Auto-derive calendar features from any datetime column
drop_date_columns	False	Drop the original datetime column after extracting its features
html	False	Save an HTML report instead of printing/showing plots
output_path	"eda_report.html"	Where to save the HTML report
numeric_strategy	"median"	"median", "mean", or "zero" — fill for missing numeric data
categorical_strategy	"mode"	"mode" or "unknown" — fill for missing categorical data
outlier_method	None	Set to "iqr" to cap numeric outliers using the 1.5×IQR rule
max_categories	10	Max categories shown in bar charts / top-value tables
show	True	Call plt.show() when html=False
verbose	True	Print the text summary when html=False

Return value

analyze() returns a dict:

{
  "df": <cleaned-or-original dataframe, plus engineered date columns>,
  "analysis": {"overview": {...}, "column_types": {...}, "columns": {...}},
  "cleaning_report": {"actions": [...]} or None,
  "feature_engineering_report": {"actions": [...], "new_columns": [...]} or None,
  "report_path": "report.html",      # only if html=True
  "figures": {...},                   # only if html=False
}