1. Data Integrity and Business Objectives

Checking data integrity is essential, but it is not enough on its own. Analysts must also ensure that the data they use aligns with the business objective. Even accurate data can be unsuitable if it does not properly support the question being asked.

Aligning data with objectives adds another layer to data integrity because every dataset has limitations that may affect analysis.

2. Matching Data to the Right Question

The process of matching data to business objectives is often straightforward:

  • If the question is about revenue, use the revenue table
  • If the question is about customer satisfaction, use customer reviews or ratings data

However, selecting the correct dataset is only the first step. Analysts must evaluate whether the data is ready and appropriate for analysis.

3. Common Data Limitations Analysts Must Address

a. Unclean Data

If data has not been properly cleaned, it cannot yet be used for analysis. Analysts must wait for or perform thorough data cleaning before proceeding.

b. Duplicate Data

Duplicate data occurs when the same entity (such as a customer) appears in multiple rows.

  • This can distort calculations like averages
  • It may appear as if there are multiple customers instead of one

To address duplicates, analysts may need to:

  • Reformat the data
  • Adjust calculation methods (e.g., averaging per unique customer)

4. Insufficient Data

Sometimes, there is simply not enough data to support an accurate analysis.

Examples:

  • Only a few months of sales data
  • Missing key tracking information

Possible responses include:

  • Changing the analysis approach
  • Finding alternative data sources
  • Creating new processes to collect additional data

Waiting for more data is often not practical, so adaptability is critical.

5. The “Incomplete Picture” Problem

A dataset can look complete because it is organized in rows and columns, but still fail to tell the full story.

Just like viewing only part of a picture:

  • Incomplete data limits understanding
  • Important context may be missing
  • Conclusions may be misleading

Analysts should never assume that a structured dataset automatically contains everything needed.

6. Real-World Example: Creating New Data

In one case, an analyst was asked to reduce the time between customer purchase and delivery for an online retailer.

Problem:

  • Existing data lacked key tracking details
  • Important journey information was missing

Solution:

  • Collaborated with data engineers
  • Created new processes to collect additional data (e.g., number of delivery stops)

Result:

  • Reduced delivery time
  • Improved customer satisfaction

This demonstrates that creating new data can be just as important as analyzing existing data.

7. Key Takeaway

Successful data analysis requires more than clean data. Analysts must:

  • Ensure data aligns with business objectives
  • Recognize and address data limitations
  • Adapt processes or create new data when necessary

Learning how to manage data issues while staying focused on the objective is a critical skill for long-term success as a data analyst.