🧠 pyedahelper - Simplify Your Exploratory Data Analysis (EDA)

pyedahelper is an educational and practical Python library designed to make Exploratory Data Analysis (EDA) simple, guided, and fast, especially for data analysts, students, and early-career data scientists who want to spend more time analyzing data and less time remembering syntax.

It's a lightweight, educational, and intelligent Python library that helps you perform Exploratory Data Analysis (EDA) faster — with guided suggestions, ready-to-use utilities, and clean visualizations.

🌟 Key Features:

• ⚡ A smart EDA cheat sheet (interactive and collapsible),
• 💬 AI-guided EDA assistant — suggests the next logical step (e.g., “View top rows with df.head()”).
• 🧩 A suite of data tools for real-world EDA tasks (loading, cleaning, feature engineering, visualization, and summaries),
• 💬 Handy code hints and examples you can copy directly into your notebook.

🌍 Why pyedahelper?

Performing EDA often involves the use of numerous syntaxes to understand the dataset, it forces the narrative that good data professionals are those who know all the Python syntaxes by heart rather than those who can interprete accurately, the output of each of the EDA steps. And more importantly, Data Analysts spend more than 80% of their analytics time on iterative EDA, some of these hours spent checking documentary and Googling stuffs.

pyedahelper solves this by combining ready-to-use functions for your data workflow, AI-powered guide with inline learning — you can see, learn, and apply the same steps.

✨ What Problem Does pyedahelper Solve?

Exploratory Data Analysis (EDA) is essential, but repetitive.

Across projects, users repeatedly:

• Forget basic pandas syntax (df.info(), df.describe(), df.groupby())

• Run the same plots without understanding what matters

• Miss data issues that affect modeling readiness

• Lose time recalling workflows rather than reasoning about data

pyedahelper addresses this by guiding users through EDA as a logical process, not a memory test.

⚙️ Installation

pip install pyedahelper

Upgrade

pip install --upgrade pyedahelper

🚀 Quick Start

import edahelper as eda
import pandas as pd

# Load your dataset
df = pd.read_csv("data.csv")

# 📚 Display the interactive EDA cheat-sheet
eda.show() -- for experienced analysts or
eda.core.show() -- for total newbies

# 🔍 Start guided suggestion
eda.next("read_csv")   # Suggests: "View first rows with df.head()"

# 💡 View an example command with short explanation
eda.core.example("describe")

From there, the assistant automatically continues:

df.head() → df.columns → df.shape → df.info() → df.describe() → ...

If you want to skip a suggestion, simply type "Next".

🔍 Modules Overview

1️⃣ EDA Guidance (AI Suggestion System)

The next() method in pyedahelper provides contextual next-step suggestions for your data analysis workflow.

Instead of remembering long commands, simply call:

eda.next("read_csv")

…and it will suggest the next logical step in your EDA, cleaning, visualization, or modeling process.

Below is a list of common helper keywords and what next() will suggest for each stage of analysis:

🔹 Basic EDA

| Keyword    | Suggestion                                                         |
| ---------- | ------------------------------------------------------------------ |
| `read_csv` | View first rows with `df.head()`                                   |
| `head`     | Check column names with `df.columns`                               |
| `columns`  | See shape (rows, columns) using `df.shape`                         |
| `shape`    | Get column data types with `df.info()`                             |
| `info`     | Summarize numeric data with `df.describe()`                        |
| `describe` | Check for missing values using `df.isnull().sum()`                 |
| `isnull`   | Get total missing values count using `df.isnull().sum()`           |
| `sum`      | Fill missing values using `df.fillna()` or drop with `df.dropna()` |

🔹 Missing Values Handling

| Keyword            | Suggestion                                                                  |
| ------------------ | --------------------------------------------------------------------------- |
| `fillna`           | Try filling missing values by data type: numeric, categorical, or datetime. |
| `fill_numeric`     | Fill numeric NaNs with `df['col'].fillna(df['col'].mean())`                 |
| `fill_categorical` | Fill categorical NaNs with `df['col'].fillna(df['col'].mode()[0])`          |
| `fill_datetime`    | Fill datetime NaNs with `df['col'].fillna(df['col'].median())`              |
| `dropna`           | Drop missing rows using `df.dropna()` if too many missing values exist.     |

🔹 Data Cleaning

| Keyword           | Suggestion                                                |
| ----------------- | --------------------------------------------------------- |
| `duplicated`      | Check for duplicate rows using `df.duplicated().sum()`    |
| `drop_duplicates` | Remove duplicates with `df.drop_duplicates(inplace=True)` |
| `replace`         | Replace wrong entries with `df.replace({'old':'new'})`    |
| `astype`          | Convert columns to proper data types using `df.astype()`  |

🔹 Visualization

| Keyword             | Suggestion                                                                                      |
| ------------------- | ----------------------------------------------------------------------------------------------- |
| `plot_distribution` | Plot column distributions using `sns.histplot(df['col'])`                                       |
| `plot_correlation`  | Visualize correlations using `sns.heatmap(df.corr())`                                           |
| `scatterplot`       | Scatter two numeric variables using `sns.scatterplot(x, y, data=df)`                            |
| `cat_num_plot`      | Use `sns.boxplot(x='Category', y='Value', data=df)` for categorical-numerical plots.            |
| `cat_cat_plot`      | Use `sns.countplot(x='Category1', hue='Category2', data=df)` for categorical-categorical plots. |
| `num_num_plot`      | Use `sns.jointplot(x='X', y='Y', data=df)` for numerical-numerical relationships.               |

🔹 Feature Engineering

| Keyword         | Suggestion                                                              |
| --------------- | ----------------------------------------------------------------------- |
| `label_encode`  | Label encode with `LabelEncoder()` for categorical columns.             |
| `onehot_encode` | Use `pd.get_dummies(df, columns=['col'])` for one-hot encoding.         |
| `scale_numeric` | Standardize numerical features using `StandardScaler().fit_transform()` |

🔹 Modeling

| Keyword                 | Suggestion                                                                |
| ----------------------- | ------------------------------------------------------------------------- |
| `train_test_split`      | Split data using `train_test_split(X, y, test_size=0.2, random_state=42)` |
| `fit_model`             | Train a model like `LogisticRegression().fit(X_train, y_train)`           |
| `predict`               | Predict outcomes with `model.predict(X_test)`                             |
| `classification_report` | Evaluate performance using `classification_report(y_test, y_pred)`        |
| `confusion_matrix`      | Plot confusion matrix with `sns.heatmap(confusion_matrix(...))`           |

This feature helps beginners and professionals alike stay productive and focused on insights rather than remembering syntax.

5️⃣ Visualization Module

Functions for exploring and visualizing data quickly.

from edahelper import visualization as vis

vis.plot_correlation(df)
vis.plot_distribution(df, "Age")
vis.scatter(df, "Age", "Income", hue="Gender")

🎨 Uses matplotlib and seaborn under the hood for fast, clean plots.

📘 The Interactive Cheat-Sheet

When you forget a syntax, simply call:

eda.show()

✨ Displays a colorful grouped guide of:

Data Loading Overview Missing Values Indexing & Grouping Visualization Feature Engineering NumPy & sklearn tips

🧑🏽‍💻 Example Workflow

import pandas as pd
import edahelper as eda
from edahelper import inspect

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

eda.next("read_csv")
df.head()

eda.next("head")
df.columns

eda.next("columns")
df.info()

inspect(df)

📦 Project Structure

edahelper/
│
├── __init__.py
├── core.py        # examples, topics, hints
├── show.py        # display utilities
├── nextstep.py   # guided workflow engine
├── inspector.py  # decision-oriented EDA checks

🛠 Requirements

Python 3.8+ pandas numpy seaborn scikit-learn matplotlib rich (for colored terminal output)

🧾 License

MIT License © 2025

Chidiebere V. Christopher Feel free to fork, contribute, or use it in your analytics workflow!

🌟 Contributing

We welcome contributions — bug fixes, new EDA tools, or notebook examples.

1. Fork the repo
2. Create your feature branch (git checkout -b feature-name)
3. Commit your changes
4. Push and open a Pull Request 🎉

🔗 Links

📦 PyPI: https://pypi.org/project/pyedahelper/ 💻 GitHub: https://github.com/93Chidiebere/pyedahelper-Python-EDA-Helper ✉️ Author: Chidiebere V. Christopher

🚀 Learn. Explore. Analyze. Faster. pyedahelper — Stop remembering syntax. Start reasoning about data.