BI and ELT: Running a data lake without a rigid ETL process

We know that building an efficient data warehouse is the guarantee for the success of a business intelligence project.

However, DW requires a super-rigorous ETL process, which takes a lot of time and effort. For this reason, many organizations are now adopting the data lake to support their BI, as it uses a much freer and more flexible method for its construction, ELT.

Want to better understand this change?

Continue reading!

Can a BI survive without a data warehouse?

The data lake tool has emerged as a promising way to deal with large volumes of structured and, mainly, unstructured data. It is a technology that allows modern companies to profoundly improve their business intelligence.

But how does this work without a data warehouse as an intermediary?

Let's explain the process!

In the data lake, data is extracted and loaded without much preparation or structuring.

Then, analysts identify the relevant data and transform it according to their analysis.

And, finally, they explore this data using business intelligence tools.

This way, the distance from extraction to analysis is shorter, saving time.

So the answer is yes, a BI survives without a data warehouse. In fact, a BI approach in a data lake represents a great victory, especially in terms of cost, time and effort savings.

All this without losing the performance and concurrency that end users demand.

But don't worry, this doesn't mean that DW is no longer necessary. Follow to find out more!

Does the data lake replace the data warehouse?

In fact, ELT (extract, load and transform) is a process that allows BI analysis, avoiding the data warehouse. Despite this, DW is not simply eliminated, much less replaced.

But, if it is possible to solve BI only with a data lake, why go to the trouble of building a data warehouse?

Simple! Because without it:

• the data will not be in a format suitable for reporting;
• data remains of low quality;
• processing will take longer and, as a result, performance will decrease;
• data will be dispersed across systems in different departments;
• historical information will be missing.

In other words, to analyze structured and more detailed business data, you need all the preparation and transformation that only a DW has. It is still used for critical business analysis on its core metrics, such as finance, CRM, ERP, among others.

For example, if management needs to see a weekly revenue dashboard or an in-depth analysis of revenue across all business units, the data needs to be organized and validated. 

This analysis example cannot be assembled from a data lake exclusively.

So why should you adopt a data lake for BI projects?

Because, effectively, “in-data-lake BI”, as this new process is being called, provides an integration that meets the demands of companies to react immediately to the contingencies of the dynamic market in which we live.

Do you not know where to start?

Our specialized team can help you. Contact us by clicking here.