User Journery Analaysis

Documentation

preprocessing_.ipynb

This file is used for preprocessing the raw data

(user_journery_raw.csv) => (clean_data.csv)

Functions Used

• group_data

  • This function aggregates user journeys based on the specified session parameter and groups by group_column (default is user_id).
  • session: Determines which part of the user journey to consider.
  • 'all': Takes the entire user journey.
  • 'all_except_last': Excludes the last entry.
  • 'Last': Takes the last specified number of entries.
  • 'first': Takes the first specified number of entries.
  • group_mask: A boolean mask used to filter rows for a specific group.
  • user_journey: Aggregates the user journey entries into a single string.
  • empty_df: Stores the aggregated results.

• remove_pages

  • This function removes specified pages (actions) from user journeys.
  • pages: A list of pages to be removed.
  • It splits each journey string into a list of pages, filters out unwanted pages, and then joins them back into a string.
  • clean_data: Applies the remove_pages function with an empty list of pages (no pages removed in this case).

• remove_page_duplicates

  • This function removes consecutive duplicate pages in each user journey.
  • It splits each journey into a list of pages, checks for consecutive duplicates, and removes them.
  • clean_data: Applies the remove_page_duplicates function to clean the DataFrame.

analysis_.ipynb

This Notebook is used for Analysis of the user_journey from the clean_data.csv which was created above

To understand the behavior of users based on their subscription types, we will analyze their journey through various pages.

Functions Used

• filter_vai_subcription

  • Purpose: To filter the dataset based on the subscription type.

  • Parameters:

    • data: The dataset to be filtered.
    • filter: The subscription type to filter by (e.g., 'Annual', 'Monthly'). Defaults to "all" which means no filtering.
    • target_column: The column in the dataset that indicates the subscription type. Defaults to 'subscription_type'.

  • Functionality:

    • If filter is set to "all", it returns the entire dataset without filtering.
    • Otherwise, it filters the dataset to include only rows where the target_column matches the specified filter value.

• split_user_journey

  • Purpose: To split the user journey strings into lists of individual pages.
  • Parameters:
    • data: The dataset containing user journeys.
    • target_column: The column that holds the user journey as a string. Defaults to 'user_journey'.
  • Functionality:
    • It splits the strings in the target_column on the "-" character, converting them into lists of pages.
    • This is useful for subsequent analysis where each page in the journey needs to be considered separately.

• page_count

  • Purpose: To count the occurrences of each page in the user journeys.
  • Parameters:
    • data: The dataset containing user journeys.
    • target_column: The column containing the user journeys as lists of pages. Defaults to 'user_journey'.
    • subcription_type: The type of subscription to filter by before counting pages. Defaults to 'all'.
    • sort: A boolean indicating whether to sort the counts in ascending order. Defaults to False (descending order).
  • Functionality:
    • It first filters the data based on the specified subcription_type.
    • Then, it flattens all user journeys into a single list of pages and counts the occurrences of each page.
    • Finally, it creates a DataFrame from these counts, optionally sorting them, and returns it.

• page_presence

  • Purpose: To count the presence of each page across all user journeys.
  • Parameters:
    • data: The dataset containing user journeys.
    • target_column: The column containing the user journeys as lists of pages. Defaults to 'user_journey'.
    • subcription_type: The type of subscription to filter by before counting page presence. Defaults to 'all'.
    • sort: A boolean indicating whether to sort the presence counts in ascending order. Defaults to False (descending order).
  • Functionality:
    • It filters the data based on the specified subcription_type.
    • Converts each user journey to a set (to avoid counting duplicate pages within a single journey).
    • Counts the presence of each page across all journeys.
    • Creates and returns a DataFrame with the counts, optionally sorted, including the total number of journeys.

• page_destination

  • Purpose: To analyze the transitions between pages in user journeys.
  • Parameters:
    • data: The dataset containing user journeys.
    • target_column: The column containing the user journeys as lists of pages. Defaults to 'user_journey'.
    • subcription_type: The type of subscription to filter by before analyzing transitions. Defaults to 'all'.
    • sort: A boolean indicating whether to sort the transitions in ascending order. Defaults to False (descending order).
  • Functionality:
    • It filters the data based on the specified subcription_type.
    • Iterates through each journey to record the transitions from one page to the next.
    • Counts how often each transition (from one page to the next) occurs.
    • Returns a dictionary where each page maps to another dictionary showing the counts of transitions to subsequent pages, optionally sorted.

• page_sequence

  • Purpose: To analyze sequences of pages of a specified length within user journeys.
  • Parameters:
    • data: The dataset containing user journeys.
    • target_column: The column containing the user journeys as lists of pages. Defaults to 'user_journey'.
    • sequence: The length of the page sequence to analyze. Defaults to 3.
    • subcription_type: The type of subscription to filter by before analyzing sequences. Defaults to 'all'.
    • sort: A boolean indicating whether to sort the sequence counts in ascending order. Defaults to False (descending order).
  • Functionality:
    • It filters the data based on the specified subcription_type.
    • Extracts all possible sequences of the specified length from each user journey.
    • Counts the occurrences of each sequence.
    • Uses a flag to ensure that each sequence is only counted once per journey.
    • Returns a sorted dictionary of these sequences and their counts.

Purpose of Creating These Functions

The purpose of these functions is to provide a detailed analysis of user behavior on a website or platform. By filtering and analyzing user journeys based on subscription types, these functions help in understanding:

• Which pages are most frequently visited (page_count).
• How often each page appears in user journeys (page_presence).
• The most common transitions between pages (page_destination).
• Frequent sequences of pages users navigate through (page_sequence).

Authors

• @ Pushpak Hinglaspure