What is a Data Pipeline?

When you hear the term “data pipeline” you might envision it quite literally as a pipe with data flowing inside of it, and at a basic level, that’s what it is. Data integration is a must for modern businesses to improve strategic decision making and to increase their competitive edge — and the critical actions that happen within data pipelines are the means to that end.

The growing need for data pipelines.

As data continues to multiply at staggering rates, enterprises are employing data pipelines to quickly unlock the power of their data and meet demands faster.

Download The Definitive Guide to Data Integration now.
Download Now

According to IDC, by 2025, 88% to 97% of the world's data will not be stored. This means in just a few years data will be collected, processed, and analyzed in memory and in real-time. That prediction is just one of the many reasons underlying the growing need for scalable data pipelines:

  • Acceleration of data processing: Time to process data is elusive; and data quality is a top concern for executives. Flawed data is everywhere; often it’s incomplete, out of date, or incorrect. In this data-driven world, spending hours using tools like Excel to fix the data is no longer an option.
  • Data engineer shortfall: Enterprises can’t stop the tide of productivity demands even in the face of a shortage of qualified data scientists, heightening the need for intuitive data pipelines to harness the data.
  • Hard to keep up with innovation: Many enterprises are being held back by a rigid legacy infrastructure and the skillsets and processes tied to it. As data continues to grow and evolve, enterprises are seeking scalable data pipelines that can easily adapt to ever-changing requirements.

The data in the pipeline.

A typical enterprise has tens of thousands of applications, databases, and other sources of information such as Excel spreadsheets and call logs — and all the information needs to be shared between these data sources. The explosion of new cloud and big data technologies have also added to the complexity of data as stakeholders’ expectations continue to grow. A data pipeline encompasses a series of actions that begins by ingesting all your raw data from any source, and rapidly transforming it into insight-ready data.

Watch Getting Started With Data Integration in the Cloud now.
Watch Now

The journey through the data pipeline.

The data pipeline encompasses the complete journey of data inside a company. The four key actions that happen to data as it goes through the pipeline are:



  1. Collect or extract raw datasets. Datasets are collections of data and can be pulled from any number of sources. The data comes in wide-ranging formats, from database tables, file names, topics (Kafka), queues (JMS), to file paths (HDFS). At this stage, there is no structure or classification of the data; it is truly a data dump, and no sense can be made from it in this form.
  2. Govern data. Once the data is collected, enterprises need to set up a discipline to organize data at scale, and this discipline is called data governance. It starts with linking the raw data to its business context so that it becomes meaningful, then taking control of its data quality and security, and fully organizing it for mass consumption.
  3. Transform data. Data transformation cleanses and changes the datasets to the correct reporting formats. Unnecessary or invalid data should be eliminated, and remaining data is enriched in accordance with a series of rules and regulations determined by your business’ needs for the data. The standards that ensure data quality and accessibility during this stage should include:
    • Standardization: Defining what data is meaningful and how it will be formatted and stored.
    • Deduplication: Reporting duplication to data stewards; excluding and/or discarding redundant data.
    • Verification: Running automated checks to compare similar information like transaction times and access records. Verification tasks further prune unusable data and can red-flag anomalies in your systems, applications, or data.
    • Sorting: Maximizing efficiency by grouping and storing items like raw data, audio, multimedia, and other objects in categories. Transformation rules will determine how each data piece is classified and where it will go next. These transformation steps pare down what was once a mass of unusable material into qualified data.
    • Share data. Now the transformed, trusted data is finally ready to be shared. People are eager to get their hands on this data, which is often output into a cloud data warehouse or endpoint application.

Download 2018 Gartner Magic Quadrant for Data Integration Tools now.
Download Now

When it comes to data processing and integration, time is a luxury that enterprises can no longer afford. The goal of every data pipeline is to integrate data to deliver actionable data to consumers as near to real-time as possible. A data pipeline should be built using a repeatable process that is capable of handling batch or streaming jobs and is compatible with the cloud or big data platform of your choice today — and in the future.

Learn more.

Talend Cloud Integration Platform delivers data quality tools to automate and simplify these processes for fast and easy data integrations. Any format, any source. Cloud Integration from Talend also includes advanced security features, 900+ connectors, and a host of data management tools to ensure that your integration runs smoothly from start to finish. Download a free trial today and let data quality be one less thing to manage.

Talend recently acquired Stitch to provide a complementary solution that will enable many more people in an organization to collect more data, which can then be governed, transformed, and shared with Talend, providing faster and better insight for all.

| Last Updated: April 3rd, 2019