Imagine you are analyzing your customer’s journey at a granular level. With event logs, you capture data on every touch point of theirs along with start time, duration, IP, etc. You also capture transactional events such as purchase data, cart history, ratings, and reviews, etc. Your analytics could include questions about your customers’ preferences, time spent in making the purchase, rating patterns and more- all split by geographies, age groups, or any other feature that is important for your organization.
To be able to answer these questions, you will have to coalesce datasets from different systems and transform them into a more useful state. You will need to map IP to geographies, derive time spent by calculating the start and end time of sessions, convert all currency values to USD, eliminate all null values, and more.
Writing individual scripts to get these enrichments right can be extremely hard. You will have to manually experiment on a sample event, or worse, deploy and wait for the event to reach the destination warehouse to review if the transformation code is fetching the right output. This makes the turn around time high and the process inefficient. Things get harder if you want to enrich nested NoSQL data.
Hevo’s Transformations make it easy for you to clean, filter, transform and enrich both structured and unstructured data on the fly through a simple Python coding interface.
With Transformations, you can load a sample event from your source with a click of a button and write quick Python transformations to clean, aggregate and enrich your data. You can even split an incoming event into multiple arbitrary events making it easy for you to normalize Nested NoSQL data.
A preview window lets you test the transformation before deploying the same ensuring that the right output is written on the destination.
Transformations on Hevo can be written on Python 2.7 environment hosted within Hevo. Additionally, the environment has all the standard python libraries enabled for you to write an array of transformations without any hassle.
Any exceptions that occur post-deployment are notified to the user over email and slack ensuring that he has the scope to take quick action on it.
With Transformations, we want to eliminate all the hardships involved in Data cleaning and enrichment. Let us know your thoughts on Transformations in the comments.