As an analyst working with HR data in systems like Workday, you may often encounter bad data—missing fields, inconsistencies, or outdated information—that can limit the insights you're able to generate. In this post, we’ll explore practical strategies to help you overcome these issues and still deliver valuable analysis, even when your data isn’t perfect.

Use Exploratory Data Analysis (EDA) to Assess Data Quality

Exploratory Data Analysis (EDA) is an essential first step in dealing with bad data. EDA helps you understand the structure of your dataset and identify any glaring issues before diving into deeper analysis. As an analyst, you can use EDA techniques to spot patterns like missing data, inconsistencies, or outliers.

For example, you can:

• Visualize Missing Data: Use a heat map or bar plot to show which fields have missing values and how frequently those gaps occur.
• Look for Inconsistent Entries: Check for variations in how data is entered. For example, if job titles are entered inconsistently across departments, you can flag those discrepancies for correction.
• Identify Outliers: Use box plots or scatter plots to identify outliers that could indicate incorrect data entry, like an employee with a salary far above the normal range for their role.

By conducting EDA, you get a clear picture of the scope of your data quality issues and can make informed decisions on how to address them. Additionally, if you spot significant issues, you can communicate these findings to your team to potentially improve data entry practices in the future.

Focus on What You Can Control

Here are some strategies you can employ:

• Prioritize Important Fields: Focus your cleaning efforts on the most critical fields for your analysis. If you're working on a project related to retention, ensure that fields like "hire date" and "exit reason" are accurate before moving forward with analysis.
• Flag Inconsistent Data: Where possible, flag inconsistent data or missing entries that could skew your analysis. For example, if a certain field has too much missing data to be useful, you might exclude it from the analysis or highlight it in your report as a limitation.‍
• Document Your Cleaning Process: As you clean the data, document what you've done and why. This helps ensure transparency and makes it easier to explain any limitations in your final report.

Communicate Data Limitations to Stakeholders

When you present your findings, be upfront about the quality of the data and how that might impact your results. For example, if there are significant gaps in certain fields, make sure to explain how those gaps could affect the reliability of your conclusions. This helps set expectations and ensures that stakeholders understand any potential flaws in the analysis.

Clear communication about data quality will help build trust with your team and make your analyses more credible.

Use Workarounds for Missing or Incomplete Data

In many cases, you'll need to work with missing or incomplete data. Here are some approaches to consider:

• Imputation: Estimate missing values based on the existing data.
• Exclusion: If the missing data is too extensive, it might be better to exclude that field from your analysis entirely. Just be sure to note this decision in your report.
• Using Proxy Data: If a key data point is missing, consider whether there are other, related fields that could serve as a reasonable proxy. For example, if exit reasons are missing, turnover rates in similar departments might provide some insight.

These techniques allow you to work with less-than-perfect data while still generating useful insights.

Reach Out for Collaboration

While your primary role as an analyst is analysis rather than data management, don't hesitate to reach out to those who manage the data to discuss any major problems you encounter. For example, if certain fields are consistently missing or entered incorrectly, flagging this for the team could lead to improvements in how data is entered moving forward.

Building good relationships with those who oversee data entry can help you advocate for better data quality in the long term, potentially leading to more accurate and insightful analyses in the future.

Conclusion: Turn Messy Data into Meaningful Insights

Bad data is a common challenge in HR analytics, but it doesn’t have to limit your impact. By understanding the reasons behind missing data, leveraging EDA, and using smart workarounds, you can turn even incomplete or inconsistent data into valuable insights. Your ability to work with imperfect data is an essential skill in people analytics, and by applying these strategies, you can still drive meaningful results for your organization.

‍

‍