Mission 2: Cleanse Data with AI

Estimated time: 10 minutes

Learning Objective

Now that you've learned how to create data pipelines, it's time to apply those skills to clean and refine a dataset. In this mission, you will perform specific actions to ensure the data is accurate, consistent, and ready for use.

Dataset

Download the required dataset: Churn.xlsx

Prerequisites

Please refer to the Navigation Guide to familiarize yourself with the platform interface.

Tip

If you're unsure about any steps, review Mission 1 first to understand the basics of the Data Transformer.

Data Cleansing Goals

We want to perform the following transformations:

"medium_of_operation" - Replace ? with Unknown
"membership_category" - Remove the word Membership
"offer_application_preference", "used_special_discount", "past_complaint" - Convert Yes/No to 1/0
"points_in_wallet" - Fill empty fields with 0

Step-by-Step Instructions

1. Import the Dataset

In the "Module 1" folder, click the "Add Item" button, select Add Data, choose Excel, and import the Churn dataset.

2. Open the Data Transformer

Right-click the Churn dataset and select Transform Data to launch the AI Data Transformer.

3. Build Your Cleansing Pipeline

Create a multi-step cleansing pipeline by entering each prompt in the search bar. You can try creating the pipeline yourself, or follow the detailed steps below.

Step 1: Replace Unknown Values

Replace placeholder characters with meaningful text:

replace the ? with the word Unknown in "medium_of_operation"

Press Enter and click Add Step to create the next transformation.

Step 2: Remove Text from Category

Clean up category names by removing redundant text:

remove the word "Membership" from "membership_category"

Press Enter and click Add Step.

Step 3: Convert Yes/No to Binary

Standardize categorical data by converting Yes/No values to binary format across multiple columns:

turn Yes and No into 1 and 0 in "offer_application_preference" "used_special_discount" and "past_complaint"

Press Enter and click Add Step.

Step 4: Fill Missing Values

Handle missing data by filling empty fields with default values:

fill empty fields with 0 in "points_in_wallet"

Press Enter.

Continue to Next Mission

Don't exit the Data Transformer dialog yet. Keep it open and continue directly to Mission 3: Export Pipelines to learn how to export your pipeline as a reusable Flow.

Visual Guide

Step 1: Open Data Transformer

Open Transformer — Right-click the Churn dataset and select "Transform Data"

Step 2: Replace Unknown Values

Step 3: Remove Text from Category

Remove Membership Text — Enter the second prompt - remove the word "Membership" from "membership_category"

Step 4: Convert Yes/No to Binary

Convert Binary — Enter the third prompt - turn Yes and No into 1 and 0 in "offer_application_preference" "used_special_discount" and "past_complaint"

Step 5: Fill Missing Values

Next Step

Keep the Data Transformer dialog open and proceed to Mission 3 to export this pipeline as a Flow.

Common Data Cleansing Tasks

Task	Example Command
Replace values	Replace "N/A" with "Unknown" in "Status"
Remove text	Remove the word "ID" from "Customer_ID"
Convert format	Turn Yes and No into 1 and 0 in "Active"
Fill missing	Fill empty fields with 0 in "Quantity"
Standardize text	Convert all text to uppercase in "Country"

Summary

You've successfully learned how to:

✓ Import datasets for data cleansing

✓ Build a multi-step data cleansing pipeline

✓ Replace placeholder values with meaningful data

✓ Remove unwanted text from columns

✓ Convert categorical values to binary format

✓ Fill missing values with defaults

✓ Save cleansing pipelines for reuse

Well done! Now let's move on to the next mission!