If you have not read my blog post on “Is Business Intelligence important for your organization?” Then head onto that page to find out that. It is essential to know what you are diving into. This blog post will take you through the basics of BI implementation. The basic 3 basic things you need to gather your thoughts around.
1. What are the various sources of data in your organization?
These are a few areas that you should think about
- Excel sheets
- Accounting Softwares
- Google ads / Google analytics and other such Google services
- Other cloud-based services that your organization uses
- IoT data
- Other Production / Payroll /manufacturing/development software that you can think of
2. How frequently is this software updated with fresh data and how frequently would you want to see data from each of these software?
3. Do you need analytics or insights which use data from two different services?
These are a few key questions that will define the architecture of your BI solution. There are a lot of other technical questions that will come into play once you get onto the actual implementation. Diving into basics of execution
- ETL – Once you have figured where your data lies. The next step is to extract the data from the source and deploy it in a central warehouse. If you are using widely used applications for your organization connecting the data from these applications is an easy job. If it is internal custom software that you have built a custom codes needs to be developed to extract the data from these applications.
As soon as this is done the next step is to transform and dump the data onto a data warehouse from where you can easily query your data.
- Data Warehouse Vs Data Lake – Depending on the type of data that you want to use for your analysis one needs to either select a Data warehouse or data lake set-up process. In a lot of cases, a combination of both can be used. If your end goal is to view a lot of unstructured data from various sources a data lake is a solution to go with. If your organization has set a pattern of collecting data like financial transactions, Sales, Operations it is best to go with a data warehouse concept.
The key to choosing the right stack is to understand what is the size of the data we are looking at and the frequency with which the data refreshes. With so many competing for products out in the market choosing the right fit has now become a role of a specialist.
- Visualization – The next and the final step is to see your queries or analysis in a graphical format. This is an essential step where you actually see the impact of the BI project. This step is commonly neglected. A lot of organizations use tools which are not user friendly. These tools pack less amount of information in one view thus makes you keep on switching tabs for the exact view that you require. This way you lose track of the actual thought. The design of the dashboard should be given equal importance. It should be monitored to customize it in the exact way a user interacts with the data. Similar to app design.