Reviewing your data

This is an note/information message.

This article assumes that you have uploaded and configured your data in to PayAnalytics, as described in the Uploading your data article.

Once your data is uploaded and your data set is properly configure, an overview page containing key visualizations and statistical overviews of your data is displayed. The purpose of the overview is two-fold:

• Ensure the quality of your data before running any analysis.

• Gain initial insights from your pay data, including your unadjusted pay pap.

This is a tip message.

This is a good time to make the final check of your data and assure yourself that it is complete and accurate.

In this article, you will learn about three main components that support data quality and insights.

You may take a look at the video belo< to explore the three components and how they might support you, or continue reading.

Investigating data quality issues: Employee Overview graph

The starting point for detecting data issues is the Employee Overview graph. Available in the Overview tab of the dataset visualization page, this component lets you examine your employee data from different angles. To see it, select the demographic variables by which you’d like to organize the data or change the variables on each axis, thereby creating different overviews of your data. Each dot represents an employee, and the color of the dot corresponds to the specific demographic value of that employee.

Initial points to check in the employee overview graph:

• Are there any salaries below minimum wage, or zero?

• Is any employee outlier pointing much higher or much lower than the rest?

Next points to check in the employee overview graph:

• Change the variables on the x-axis. Do you detect any abnormalities across different groupings?

• Do you have currency support? Do you see any indications that calculations are off?

To investigate any abnormalities:

• Hover your mouse over the dot(s) to view further details; or

• Click and pin the employees for review in the employee list below.

This is a tip message.

By default, the employee graph orders the groups based on increasing average compensation (or the y-axis variable), but you can choose to sort them alphabetically in the graph's settings. You can also choose different graph types that summarize the compensation distribution.

Investigating data quality issues: Employee List

To better investigate why a dot representing an employee is an outlier on the employee graph, consider the additional data fields across the columns. You can click on the dots in the employee graph to pin the data of any employee of your choosing at the top of the list.

The employee list allows you to filter and sort the data by data categories or numerical values from low to high (or reverse). When applying a filter to the employee list, the filter automatically applies to the employee overview graph and the summary statistics. You can also hide or show as many columns as you would like to view, from your dataset.

This is a tip message.

You can also leave comments within the employee list. This might be useful if you have discovered an explanation why a specific employee's compensation is an outlier. Comments will be visible to your team members and across data sets, to avoid having to re-investigate the same issues multiple times.

Gaining initial insights: data summary and summary statistics

Once you have confirmed that your data set quality is in good order, explore your summary statistics. These statistics are based entirely on your data set, even before any pay equity analysis is performed.

In the data summary graph and summary statistics data table, you will find:

• Distribution (percentiles and average) of your employee data

• Compensation across your selected variable (Note: Y-axis in the employee overview graph, is now X-axis in the data summary graph)

• Demographic groups (Y-axis), for example gender, ethnicity, or a combination of both

One of the most important insights to gain from the data, is your unadjusted pay gap, which can be found in the table as the percentage difference of a demographic group compared to the reference demographic value (which automatically would be the highest valued group).

For example, if your demographic variable is gender, and male employees are your higher paid reference group, the percentage difference of female average pay, compared to male average pay, is your unadjusted pay gap.

While the unadjusted pay gap is an interesting starting point for exploring pay differences, there are oftentimes explanations for why one demographic group is paid more than others, which does not relate directly to their demographic value.