The data quality widget lets you know how much of your data is using "usage data" vs estimated data. As a best practice you always want to include as much primary data as possible in order to ensure that you are getting an accurate understanding of your emissions.
For example: If you have a natural gas emission source and you enter an estimate based on the amount you spent from an accounting report, that will not be as accurate as entering in the amount of fuel used.
You can view your data quality based on the % of emissions or % of data points to see how material these estimates are.