Brook Preloader


Factors To Consider For Building A Data Lake


Before going to know what a data lake is, let’s understand different types of data and how these fit into the concept of a data lake.

There are essentially 4 types of data – structured, unstructured, semi-structured and binary.

  • Structured data is the data that resides in the tables of a relational database.
  • Unstructured data is the data which is present in the form of documents like PDFs and emails. When we say unstructured data, it needs to be understood that the information inside these documents is not in any particular format. We cannot modify the content in these formats.
  • Semi-structured data is the data that resides in csv, xml or JSON files. These formats need to be converted in a way that the database can read it.
  • Binary format comprises of images, audio and video files. This is basically an oral or visual representation of data.

Data lakes store various types of data into a single centralized repository and generate the required reports, which help in making key business decisions. Now-a-days, data is represented in visual format to understand easily with some advanced concepts such as machine learning and advanced analytics. In a nutshell, a data lake is something which stores data and generates appropriate reports for making wise business decisions with advanced analytics and visualization techniques in a single environment.

Figure 1: Simple way of representing a data lake


Below image depicts the process of merging various sources into a data lake, making transformations and performing the actions on data such as collecting data insights, advanced analytics, machine learning and so on.

Figure 2: Architecture of a data lake

Several things need to be considered to build a data lake, which are discussed as follows:

  • Data Ingestion – Pulling data from various data sources and loading into the data lake with the help of connectors.
  • Data Retention – As the unused data lies in the data lake, care should be taken while defining data retention policy.
  • Data Quality – If you have quality data in the data lake, it will help to acquire good insights for business.
  • Data Storage – When it comes to storage capacity, data lakes are easily scalable and should be integrated when the data increases.
  • Data Discovery – Data lakes have a centralized repository, which allows us to get better actions on self-service analytics.
  • Data Auditing – As soon as the data begins to load into the data lake, it needs to be tracked and changes need to be captured by the users with respect to time.
  • Data Governance – Data lakes tend to be highly available and can easily be integrated into an organization within various departments.
  • Security – The data must be protected at all times and has to be integrated with only the authorized users.
  • Data Processing – The processing of data lakes is very responsive and cost-effective.

Many other factors contribute to the fact that a data lake is preferred over a data warehouse or database.


Data lakes are very cost-effective in terms of scalability and flexibility so eventually, they reduce costs in the long run. They are easily adaptable for new changes when needed. With the advantage of centralization from different sources, users from any department can have access to the data lake without any conflicts. There is no limitation for loading data into data lakes. They can store unlimited data with excellent data processing.

By monitoring the unused or unusable data, we can control or restrict these types of issues as it will impact the performance when it is expanding in size. One should be attentive and careful in this aspect while designing the data lake.

This article is a simple overview of a data lake. There are certain cloud providers such as Azure, AWS, IBM, Google Cloud etc. providing customized data lakes for their customers. For example, Azure Data Lake runs on Hadoop-based storage system (HDFS), so it can vary from one cloud provider to another.

Contact for further details

Abdul Rahiman
Sr. Specialist – Data Analytics Visualization

0 0 vote
Inline Feedbacks
View all comments