AutoEDA with glook

Hi friends,

Welcome! I’m excited to discuss a crucial and mandatory step in data analytics: Exploratory Data Analysis (EDA) and the Automation of EDA process. The continuation of my previous post (https://www.dhirubhai.net/feed/update/urn:li:activity:7208326496350818305/).

?

What is EDA?

The Data analysis is performed using the standard data processing methodology: CRISP-ML(Q) - the CRoss-Industry Standard Process for Machine Learning with Quality, also known as the machine learning framework.

In the second stage of CRISP-ML(Q) methodology, we conduct preliminary analysis through standard steps called as EDA – Exploratory Data Analysis.

Exploratory Data Analysis (EDA) - We start with univariate analysis, where we examine each column in the dataset individually. This includes:

1. Measures of Central Tendency: Calculating the mean (average), median (robust to extreme values), and mode (most frequent value).
2. Measures of Dispersion: Assessing variance, standard deviation, and range to understand data spread.
3. Skewness: Measuring asymmetry.
4. Kurtosis: Examining the data's peakedness relative to the center.

These calculations, along with visualizations like bar plots, histograms, and box plots, help us understand our data better.

However, as datasets grow larger, the manual EDA process becomes iterative and time-consuming. To address this, researchers have developed various AutoEDA libraries, including Sweetviz, Autoviz, D-tale, and Pandas Profiling, etc.

I’m proud to highlight our AutoEDA library ‘glook’ that @Gaurang and I have developed.

‘glook’ is a completely interactive, no-code tool with a graphical interface (GUI). It offers features like:

• General Data Insights
• Correlation Coefficient Heat Map
• Numerical and Categorical Data Overview

Univariate Analysis:

• Statistical Calculations (for numeric and categorical data)
• Visualizations: Histograms, Box plots, Q-Q plot

Bivariate Analysis:

• Scatter plots, Line plots, Bar plots, Box plots, Violin plots, Strip charts, Density contours, Density heatmaps, Polar plots

Tri-variate Analysis:

• 3D Scatter plots, Distplot

And many more.

For more details, check out the medium article by @Gaurang.

Stay tuned for more updates as ‘glook’ evolves into an AutoML library (pre-built version already available with unsupervised learning techniques).

I request and encourage all, to explore the ‘glook’ library to make EDA process simple and fun.

If you happen to find any bugs please report them, that will help us to enhance the package for a better community usage.

We would love to hear your experience on glook.

https://pypi.org/project/glook

#DataAnalytics #EDA #MachineLearning #AutoEDA #DataScience #DataVisualization #glook