Exploratory data analysis is a popular approach to analyse data sets and visually present your findings. It helps provide maximum insights into the data set and structure. This identifies exploratory data analysis as a technique to understand the various aspects of data.
For one to better understand the data one must ensure that the data is clean, has no redundancy, missing values, or even NULL values.

Types of Exploratory Data Analysis

There are three main types:
Univariate: This is where you look at one variable (column) at any single time. It helps one understand more about the variable’s nature and is termed as the easiest type of EDA.

Bivariate: This is where one looks at two variables together. It helps one understand the relationship between variables A and B whether they are independent or correlated.

Multivariate: This involves looking at three or more variables at a time. It is identified as an “advanced” bivariate.

Methods

Graphical: This involves exploring data through visual representations such as graphs and charts. Common visualisations include box plots, bar graphs, scatter plots and heat maps.

Non-graphical: This is done through statistical techniques. Metrics used include mean, median, mode, standard deviation and percentiles.

Exploratory Data Analysis Tools

Some of the most common tools used for EDA include
Python: An object oriented programming language used to connect existing components and identify missing values

R: An open source programming language used in statistical computing

Steps

1. Understand the data - See what type of data you are working with; number of columns, rows, and data types.
2. Clean the data – this involves working on irregularities like missing values, missing rows, and NULL values.
3. Analysis – Analyse the relationship between variables.

Sample EDA using Python

The dataset in use for this example is the Iris data set - available here

1. Load the data using the pandas library. ```

df = pd.read_csv(io.BytesIO(uploaded['Iris.csv']))
df.head()

![Image description](https://dev-to-uploads.s3.amazonaws.com/uploads/articles/jc0v1bsat7wwso2fwvow.png)

2. Identify data types
`df.info()`

![Image description](https://dev-to-uploads.s3.amazonaws.com/uploads/articles/u72xoee9g1viirjge81r.png)

3. Clean data e.g. checking for NULL values
`df.isnull().sum()`

![Image description](https://dev-to-uploads.s3.amazonaws.com/uploads/articles/7b1mg3m5brmegcahh7dt.png)

4. Non-graphical analysis of the data to give variable info
`df.describe()`

![Image description](https://dev-to-uploads.s3.amazonaws.com/uploads/articles/fc0ln6hh5g64c428qbj3.png)

5. Graphical analysis to show variable correlation or independence

df.plot(kind='scatter', x='SepalLengthCm', y='SepalWidthCm') ;
plt.show()

![Image description](https://dev-to-uploads.s3.amazonaws.com/uploads/articles/msztylnnkbxfqw41j98c.png)