Transform Your Data in Real-Time With Datazoom

Although post-processing data is often a necessity, it can create challenges as well. Whether it’s scrubbing data after it comes in, connecting data sets together, or other activities, post-processing can take valuable time, introduce errors, and increase the time between receiving data and being able to use it to make business decisions.


Are You Spending Too Much Time Processing Your Data When It Arrives?

Any Delay With Your Data Can Impact Your Business

Most of the time, when we talk about data delay, we are referring to the delivery speed. For example, in streaming a live event, receiving data even a few seconds behind an error can result in subscriber churn. But there’s another delay to being able to use data quickly: post-processing. Raw data, especially when it’s coming fast and furious from multiple sources, needs to be processed before it can be analyzed and employed for business decisions. But that processing can be very time intensive. What’s more, post-processing (whether it’s manual or automated) can introduce errors into the data that is eventually fed into dashboards and other analytics tools.

Process Your Data At The Time of Acquisition to Avoid Errors and Reduce Time.

Data Transformations: Post-Processing in Real-Time

Transformation Explained

Datazoom’s data transformation feature utilizes a filter-based approach to identify data that qualifies for modification. Users specify business rules against which incoming data is evaluated as it passes through the filter. If a match is achieved, the data is modified in one of three ways: adding a new key value pair (derive), changing a value based on conditions (convert), or renaming a datapoint key or message node location (rename).

The Technology Of Transformation

Different Transformations For Different Business Needs

Customers can build business rules in their Datazoom data pipeline configurations to handle transformations in a variety of different ways. Once a business rule is triggered by the data filter (upon identification of a data point that is flagged for transformation), the appropriate transformation is carried out in real-time so that the processed data is delivered to the appropriate place.

  • Rename

    Datapoints can be renamed. For example, to bring different datasets into compliance around a specific unique key (i.e., a user ID) where the key is named differently. This ensures joins and other SQL operations can be carried out without any extra work.

  • Convert

    The conversion transformation allows you to change a datatype (such as a string value to a boolean value) or the value of a data point. For example, num_errors_ads may have a value of "none" but you want it to be an integer of "0".

  • Derive

    Derivation encompasses if-then situations to carry out a specific action against a datapoint. For example, if "os_name: roku", add "device_platform: set_top_box". This carries out an automatic action to facilitate faster use of data once it hits the dashboard or datalake.

Real-Time Transformations Reduce Your Post-Processing Time

Use Your Data Faster. Guarantee Correctness.

Reduce Post-Processing Time

Because the transformations you might normally carry out once the data arrives in your datalake happen in real-time, you save time on manipulating data which allows you to act quicker.

Make Faster Business Decisions

Because the post-processing is done in real-time, you can make use of your data quicker when it arrives. This is especially relevant when the data concerns real-time events.

Automation Ensures Less Errors

When you build business rules (which can be carefully vetted) that are applied in a real-time automated method, you reduce the margin of processing errors, ensuring the data you use to make decisions is correct.

Scroll to Top