What Does Dimensions Means?
Dimensions consists of various variables that can be used to characterize and categorize data into an organized context of data. The dimensions provide a framework for organizing and analyzing data as each dimension gives a distinctive perspective to the dataset that aids in understanding the data at hand. In other words, dimensions are ways of slicing information which stored in terms of measures to perceive complex datasets from multiple angles. Let’s consider a furniture business that sells chairs, the dimensions would be product performance by regions, by population, and by time. The measures help to add a meaning to the dimensions such as quantity, costs, and revenue. For example, company XYZ has total revenue from selling chairs of $1M, total costs 600K, and total quantity 80K units. By organizing the business data into various dimensions such as regional, Southeastern region for example, the company can gain insights about which region has the highest revenue, lowest costs associated with acquiring the product, and highest demand for its products.
What Does Attributes Means?
Attributes on the other hand, gives a deeper understanding of the dimensions by describing the characteristics of specific measures and values. These attributes can further slice the data into other filters such as subcategories and colors to build a relationship between different measures. The relationship can be either one-to-one, one-to-many, and many-to-many. Considering the same XYZ company that sells chairs, the attributes can be the colors of the chair, chair sizes, and chair styles. For example, the company has a revenue from selling chairs of $1M, total costs 600K, and total quantity 80K units. By slicing the regional dimension data into attributes, the company can gain insights that, Southeastern region for instance, has highest quantity demand for large White balcony chair, lowest costs associated with acquiring/manufacturing the products, and highest revenue. Therefore, Company XYZ may want to concentrate on Southeastern region by increasing the production and availability of its large White balcony chair products to maximize its revenue and market share in Southeastern region. The relationship with here is many-to-many meaning that many large White balcony chairs are being sold to many costumers in Southeastern region.
What Does Hierarchies Means?
Hierarchy is a structured grouping of data elements or variables into more specific levels. As it usually related to attributes, it entails classifying data attributes according to their relationships and placing them in a hierarchical order. Data analysts can navigate and analyze data using hierarchies at various levels of detail down to more specific subsets. Considering the same XYZ company that sells chairs, the hierarchies can be the states, population, and time. For example, the company has a revenue from selling chairs of $1M, total costs 600K, and total quantity 80K units. By utilizing the data from the previous examples that the company has highest quantity demand, revenue, and lowest costs in Southeastern region of its large White balcony chair products. The company can gain further insights about which state in the Southeastern region has highest revenue, highest season demand, and highest market share based on population. As an illustration, Florida for instance, has highest revenue in Southeastern region, Miami has highest demand in Summer, and Miami has highest market share in Florida with respect to large White balcony chair products. Hence, the company may want to concentrate on Miami during the Summer season to maximize its sales and revenue. The relationship of the hierarchies should always be one-to-many relationship to ensure that dataset is clear and not misrepresented. Think of this relationship as a “tree” that has branches. For example, Florida is the tree, and the cities are the branches. Or the year of 2022 is the tree, and the seasons and months are the branches. This relation helps to represent the data in a clearer way and link information together where they should be to help the analyzing the data.