Data modeling, what is it and why does it matter? One could argue that data modeling is the most important part of any report or dashboard. A report or dashboard can have a lot of compelling visuals and tables, but if the data is not correct then it is just a pretty picture that really does not convey any useful information. Much more concerning is that it is telling an incorrect or incomplete story. Either way, it’s not performing its intended function which is providing accurate and insightful data for key business decisions. Keep reading to learn about and how to get the most out of business reporting using these 3 Power BI Modeling Tips.
Microsoft Power BI (Business Intelligence) incorporates a very useful tool for ensuring that we can model the data and make sure we build on meaningful information. Microsoft’s Power Query allows the report developer to connect to the data source, import the data, and model it. Knowing how visual dashboards are created can help understand the importance of Data Modeling in Power BI.
So, what is “Modeling the Data”? Modeling refers to building a data model that a report or dashboard will then source its information. Data modeling is the practice of cleaning the data, and naming columns and queries in terms of easily understood conventions. The report developer needs to be able to clearly see what each column is and how it will relate to the report being developed. This phase is arguably the most important step in the process. If the Data Model is not clean and consistent then this translates to a report that is confusing to build and use.
There are some important things to keep in mind when modeling data.
- As queries are added to the data model, it is important to name them. As the data model grows it will be very important to know what each query is and what data it processes. A meaningful name such as “customer list” is much more useful than the default “query 1”.
- Next, only keep the data to build the report you need. One may import a table that has 100 columns of data but only 10 of those columns are relevant to the report or dashboard being developed. If the report only needs ten of the columns, then having the other 90 columns in the data model will add overhead and consume more resources. The smaller the data model, the better the performance will be so, only kept what is needed. Data can always be added to a model if it is required.
- Name the steps that are taken when changes are made in the data model. Power Query has a list of applied steps as changes are made, this of course defaults to step 1, step 2, and so on. It will be very difficult to remember all the steps taken in the modeling process without naming each step with a meaningful name. So, if a column is renamed in a step, then name the step “renamed column x to y”. This will prove very useful as the modeling process takes place.
For further information on Microsoft Power BI, please see the below information links:
About Microsoft Power BI
Power BI Tips and Tricks
Schedule a Free Microsoft Power BI Demo
Posted by iCepts Technology Group, Inc. a Microsoft Cloud Solution Provider in Pennsylvania