Performed exploratory data analysis on covid tweets collected from 25-07-2020, with #covid19 involved.

[errpv78]

Approach
Opening the columns.txt and understanding the columns their types and description.
Loading the dataset and viewing its sample, dimensions, and a sample row.
Checking Frequency distribution and description in integer columns like user_followers, user_friends.
Checking for null and missing values and cleaning the data.
Exploring the unique values in columns their frequency and maximum lengths of columns to understand more about the distribution of data.
Filtering columns to get relevant information needed for better grouping and understanding data like adding tweet_date for date of tweet column from the date column which had both date and time to look for the frequency distribution of tweets date wise.
Plotting tweet lengths to see a variation of tweet lengths in words.
Building and training sentimental analysis Bert model, the link for separate training file included below.
Preprocessing the tweet text to filter out unnecessary words and characters and adding sentiment column to the data frame and saving data frame to a new CSV file.
Checking overall sentiment distribution among tweets.
Subsetting data frame with conditions on columns to understand distribution among columns.
Visualizing sentiments among top values and frequency distribution of sentiments with respect to other columns.
Exploring hashtags column to understand different hashtags and their relation with sentiments and other columns.