Data Analysis/Data Warehousing
The purpose of Data warehousing is to extract data from multiple data sources then transform that data for better analytical processing. Data warehouses are designed for efficient reading compared to writing data. In this process, the data is categorized as dimensions and facts.
So what is a fact? Generally Facts are numeric values. Price, Cost, Quantity, Marks are examples of a fact. Aggregating a fact helps in measuring performance of a business/student/process etc. So an aggregated fact is a measure. Sales, Margin, Total Marks are examples of measures. However measures alone are not meaningful or they does not provide complete picture in most of the cases. For example, if somebody says my business sales is 1000, that raises many questions like is it for a day/week/month, are those us dollars or indian rupees etc. Here day, week, month, us dollars, indian rupees are attributes which give some context/meaning to measures. Some attributes can be grouped under a bucket called dimension. Day, week or month can be grouped as Time dimension. Product Category, Sub-category, product brand, Sub-brand can be grouped as Product dimension. However in general Dimension and Attribute are treated as synonyms by most people.
To summarize,
Measure - How you measure something or someone. (Example: I want to measure a business performance by Sales)
Dimension - By which you measure something or someone. (Example: I want to measure Sales by Day by Branch by Sales person).
- Kranthikumar Miryala, 9+ years of experience in BI
Measure is a value on which some sort of mathematical function can be performed. A good example of a measure is revenue of a company. You could add revenue, you could average revenue. Just looking at revenue is useful. What if one wants to know revenue per Day or Revenue per Product or Revenue per Country? These are dimensions : Date, Product, Country. Dimensions give you context for a measure. An attribute is similar to a dimension and sometimes is used interchangeably. However, one dimension can have multiple attributes. For example, take the date dimension. Every date value has multiple attributes : What day of the year is a date? What week of the year is a date?
- Siva Kolappa , Be Happy, Healthy, Kind & Curious!
https://www.youtube.com/watch?v=cwpL-3rkRYQ&feature=youtu.be
-Rama Kattunga , Data in Lakes or Warehouses. Data Management,
