The higher the data quality score is, the better the predictions that IBM Watson Analytics can provide.

If the quality of your data is low, the accuracy of the analyses in your explorations and predictions is less reliable.

Fortunately, you can improve the quality of your data with IBM Watson Analytics.

When data is loaded, Watson Analytics will read and analyze the data and determine a data quality score that describes the data's ability for making predictions. The higher the score, the better the data quality. If you provide a high-quality dataset, Watson Analytics provides a high data quality score.

You are able to see the score that is associated with each dataset in the list of assets on the Welcome page. As an example, a score of 68 indicates a dataset of medium quality. The lower the score, the higher the number of outliers or missing values and other issues.

To obtain a higher (data quality) score, clean your data (before you load it into Watson Analytics) by doing the following:

• Remove summary rows and columns from your data file
• Eliminate nested column headings and nested row headings

Load your data into Watson Analytics. Review the data quality score that is given to your dataset. If your data quality score is not satisfactory, repeat the cleaning process.

For more information, see Optimizing the quality and usage of your data within the online product documentation (https://watson.analytics.ibmcloud.com).