This Jupyter Notebook focuses on preprocessing and visualizing data from an Instagram profiles dataset. It includes data loading, inspection, visualization, and some data preprocessing steps.
- The script loads a dataset from a CSV file named 'Instagram Profiles - Github Hashtag - instagram_profile.csv' into a Pandas DataFrame named
df. - It displays the first few rows and provides information about the dataset using
df.head()anddf.info().
- The script defines a Python class named
visual_preprocessto encapsulate data visualization and preprocessing functions.
- The class contains various methods for exploring and preprocessing the data:
_row_col: Helper function to calculate the number of rows and columns in the DataFrame.disp_tot_row_col: Displays the total row and column count.missingv: Visualizes missing values using a heatmap._null_calculator: Helper function to calculate the percentage of null values in columns.null_percentage: Calculates and displays columns with a specified percentage of null values.get_col_empty: Returns columns with null values above a specified threshold.
- Columns with a high percentage of null values (above 50%) are dropped from the DataFrame.
- Various data visualizations are created using Seaborn and Matplotlib, including:
- Distribution of 'posts_count' using a histogram.
- Filtering and exploration of records with 'posts_count' greater than 2000.
- Scatterplots of various features ('followers', 'following', 'highlights_count', etc.) with respect to different account types and privacy settings.
- Bar plots showing relationships between 'is_business_account' and 'is_professional_account' with 'followers' and 'following'.
- Additional scatterplots exploring features related to 'following' and 'followers'.
- The script defines a function (
hashtag_freq) to extract and analyze hashtags from the 'post_hashtags' column. - The function counts the frequency of hashtags and displays the top 10 most frequently used hashtags in the dataset.