Unified Analytics

Prism presents every data source it connects to in a unified way, called a Dimensional Model.
The main benefit of this model is that it allows you to look at your data in terms of business entities,
rather than rows and columns in a table. Each unique value within a field becomes a distinct member
within the field’s corresponding dimension. This makes it simple to navigate through large amounts of
data, quickly identify the members you are looking for and start your analysis from there.

The Dimensional Model also allows for easier and more elaborate time analysis. Typically, fields that contain time information
appear in the form of time stamps (an exact time, down to the millisecond). This makes aggregated analysis very difficult.
Prism automatically turns time stamp fields to dimensions that contain different levels of aggregation (Years, Quarters, Months,
Days, Days of Week, etc). Performing analysis at any level of time is achieved instantly and with no extra work.

Finally, this model is exactly the same in nature regardless of the data source behind it – so using Prism over any type of data is
done precisely the same way. A spreadsheet containing sales information for customers and products will look and behave exactly the same
as a similar table in a database table within the Prism environment.

Multidimensional Filters

A powerful feature of Prism’s Dimensional Model is that it can be filtered in a multidimensional way.
In simple terms, instead of filtering out rows from a table, you filter each dimension in its own context,
or in the context of other dimensions.

Imagine a table with 4 columns – Customer, Product, Time Of Purchase and Sales Amount. Each row describes the purchase of a
single product by a customer and the value of the transaction. Now think how difficult it would be to extract a list containing the
5 customers who account for the most total sales together with the products that account for 80% of each customer’s purchases.

Whether you do this in a spreadsheet or over a database – this is no simple task, and out of most users scope of technical
capabilities.

With Prism, using its dimensional model, this can be done in seconds. First, you’d apply a Top Ranking filter over the Customers
dimension and tell it to return 5 members, then you’d apply a Top Percentile filter on the Products dimension and point it to 80%.
Dropping both these filters on a Pivot widget will return exactly what we are looking for.

Examples of filtering options:
- Ranking Filters (best/worst products in terms of sales)
- Percentile Filters (products that account for 80% of total sales)
- Label Filters (products whose serial number begins with x-xxx-x)
- Value Filters (products whose sales in USA are greater than 1000 USD)

Multidimensional Formulas

Most tables contain values that are meant to be measured in some way - the value of a sales transaction,
the cost of each product in a sales transaction, the number of products sold, etc. Multidimensional Formulas
are a powerful tool for extracting more information from you existing data.

Sometimes, your measurements require you to modify the raw values in the table – like converting to a different currency if
that raw data is in dollars and your report needs to be in Euros. Other times, the measurements you require do not inherently
exist in the source data all together.

Prism lets you create Calculated Measures that act as virtual fields in which each value is evaluated using a formula. Converting
currency, for example, is achieved by the following formula: [Sales in USD] * [USD to Euro Rate].
Learn more about currency conversion.

Multidimensional formulas can express more than simple arithmetic. You could, for example create a measurement called “Sales in 2008”
by specifying the formula ([Sales], [Time].[2008]). Combining this measurement with the Customer dimension will return the total sales for each customer in 2008.

Finally, you can apply aggregation functions over filters. If we created a filter called “Products sold more than twice”, creating a
formula Count([Products sold more than twice]) and combining it with the Customer dimension will return the number of
products each customer bought more than twice.

Porting From One Database to Another

The Prism Unified Dimensional Model is exactly the same in nature regardless of the data source
behind it – so using Prism over any type of data is done precisely the same way.

Prism documents created over one type of source can be easily ported to work against a different
data source with the same inherent structure.