What is Granularity?
Granularity is the level of detail of your data set. The level of granularity of your data will determine the accuracy of your analysis.
Lower level of granularity provides a more aggregate view, while a higher level of detail provides a more specific view.
In the EU Superstore data set for example the higher granularity and lower aggregation is the Row ID, whereas Category is considered the higher aggregation and lower granularity.
How to identify the granularity of your data?
You need to examine the data itself to check the level and precision of your data. The higher level of detail of the data can be set by one or multiple columns and will be determined by a unique value for each row.
Simple Granularity
You can verify that Row ID has a unique value for a single row (EU Superstore data).
Multiple Columns Granularity
Product ID and Name become the unique values for product information, when we exclude Row ID for example. This turns Product ID and Name as the higher granularity in this example.
Using the Tableau to identify your data
Go to the data source and check how many rows your dataset has. In this example, it is indicated that there are 9994 rows in this dataset.
The next step is access the worksheet, add count Order ID to the columns and start dragging the measure names to the rows. By analyzing the data before, there is an indication that Product ID, Customer ID, Order ID or Row ID could possibly be your highest granular data.
In this first example, there are many products IDs for the same row (blue) and the number of rows (yellow) is 1862.
Adding Customer ID (see below) the total number of rows is 9944. It looks like there is one single value for each row, but the number of rows (9944) is not equal to the one in the data source (9994).
We can verify there are 50 rows with two values after the application of the filter. This is another indication that the Customer ID is not the higher granularity for this data.
The addition of Order ID increase the number of rows (9986). However, there are still 8 rows with value equal to two. The information generated still does not reveal the higher granularity of the data.
We can verify that the total number of rows matches the number of rows of the data source (9994) after add Rows ID. There are no values higher than 1 for each row after filtering.
This is an indication that Row ID is the higher granularity for this dataset. Be mindful of data with multiple columns granularity. This tip is not recommended in this case.
Conclusion
The granularity of your data should be identified before starting any analysis. It is important to determine the level of granularity, so the accuracy of your analysis is associated with its goals.
In addition, Tableau could be a tool to analyze the data itself and determine the granularity.