👌🏾 The Data Quality Report

The Data Quality Report helps you make sure your model is getting clean and useful data to give you its most accurate results. When your model refreshes, Recast will check your data for any issues and if any are found, these will be logged in the Data Quality Report.

The report contains:

• Count and description of the issues found with your data
• Changes in data between prior refreshes and this refresh
• The date the report was generated, when the data was last refreshed, what date the current refreshes' data is through, and what date the prior refreshes data was through. This will help you determine which dataset time period the model refreshed on.

If your data for the latest refresh has any issues, you will receive an email informing you of the issues so you can make any necessary checks or corrections to your data before your model run.

❗️

The Data Quality Report is only created if your data has issues

How to use the Data Quality Report?

Click the link in your email to view your data validation dashboard. Alternatively navigate to the ‘Data Quality Report’ tab in the Recast App.

Under Data Quality Report click the tab ‘current model’ to view the data quality report for the model you currently see on your dashboard. To view the Data Quality Report for the most recent data ingest and upcoming refresh, click the tab ‘latest’

• Use the ‘current model’ report to contextualize the results you see on your dashboard
• Use the ‘latest’ report to catch data issues before the model run

At the top of the page, you will see the date the report was generated, the model name that the report references, and any warnings or errors in summary.

These warnings provide insight into the patterns in the data that may make it difficult for the model to provide the best results possible. Recast recommends you flag this with your internal point of contact for data to resolve any errors or warnings.

For each report, you'll also see a breakdown of when the data was last refreshed, what date the current refresh's data is through, and what date the prior refresh's' data was through. Based on this information, you can determine the data set on which your model was refreshed on.

Finally, you'll see a breakdown of any significant changes in data between prior refreshes and this refresh.

As indicated in the quality report, significant changes in data can lead to model instability and are worth reviewing to ensure there were no data quality issues that should be resolved.

Common Warnings and Errors

Here are all of the possible data issues the report can describe.

💡

Issues are Warnings unless marked Critical. Critical errors pause model runs.

• Historical revisions detected in channel spend*
• Channels go to zero in the last week
• Unusually low KPI values
• Missing data for certain dates 🚨(Critical)
• Duplicate dates detected 🚨(Critical)
• Historical revisions detected in KPI
• Recent data is not newer than historical data🚨(Critical)
• Less data now than was in prior data

*The data table will report ALL historical changes for reference, but only changes >$1000 trigger the warning.

🚨 downgraded to warning if we manually relaunch

How to resolve data issues

🚧

The report will point out problematic data points in the included tables.

Ex. if dates are missing, we indicate which dates are missing.

• Check the problematic data points indicated in the table
• Check to make sure you uploaded a complete dataset
• Dig into why you may be seeing historical revisions by looking for data entry errors or delays in when data comes from your vendors.

Data Quality Email

Every week when your model refreshes, you will receive an email to let you know if there are any issues with data quality. See below for an example of what the email will look like. Clicking "Click here" in the email will take you to the full Data Quality Report in the Recast platform.