It’s an exciting time to get started with data in Einstein Analytics. The Einstein Analytics product team has worked very hard making it easy for users to prep their data ready for building amazing dashboards and predictions.
In Summer ’20, we are seeing the first results of their work with the public beta of the new data prep platform. I’m looking forward to showcasing the best of Einstein Analytics Data Prep and what you can expect in Summer ’20 and beyond – but first, let’s find out why it was time for a change.
Why Is It Time for a Change?
Back in 2018, user research showed that preparing data was the most challenging and frustrating part of using Einstein Analytics, which is an essential part of using the platform. It also showed that new users took up to 6 months to be successful with Dataflows. After receiving the feedback, the Salesforce product team acknowledged that it had to be changed, and embarked on a journey to make data prep effortless. This resulted in a lot of research on usage and design considerations.
Einstein Analytics Data Prep – The Benefits
Embarking on this journey, the team didn’t just want to make prepping data more intuitive, though user friendliness was definitely key to get this product right. They wanted to make it more powerful. An output of the user research defined the pain points users had working with the dataflow and recipes. This generated a lot of opportunity for ideas like data preview, orchestration of flows, connectors, smart recommendations and much more (which I cover in detail in this blog post). The main goal was to deliver an end to end solution, effortlessly bringing data in, prepping the data with ease, and deliver the data ready for reporting.
Another key opportunity, was to leverage Machine Learning (ML) to make the data prep more powerful, and you will see the ML capabilities in the Summer ‘20 beta. Working with the new data prep, you are able to predict missing data to get more complete values and detect sentiment of unstructured data to be able to manage comment fields better. This is just the start and as new releases come so will more ML capabilities.
From a Salesforce admin perspective this is very powerful. 1) You don’t have to be a data scientist with model building experience to leverage the ML capabilities 2) You don’t have to extract your data from Salesforce, train a model to work on your data, apply it and import it back into Salesforce. The heavy lifting has been done for you. If you want to understand more of how ML is influencing the new data platform, check out Jim Pan’s (Product Manager on the data platform) blog on it: New Data Prep platform made easier with native machine learning.
What’s Coming Up?
Perhaps the most noticeable that is being released is the improved user interface, something that looks similar to other tools you know from the Salesforce platform like flow. When you are prepping your data you have a nice overview with clear icons and colors that represent the different nodes. On top of that, you can click on any node and see a preview of the data at that stage. Something I believe every experienced dataflow builders will be excited to have. But there are many other cool features to look forward to:
- Be able to join data sources using not only lookup but also joins (left, right, full, inner) as well as append.
- Transform nodes with quick transformations like reformat date, calculate fields etc.
- ML capabilities like predict missing values and detect sentiment
Make sure to check out the release notes for the details on all the features that are available in Beta.
As data prep will be in beta in Summer ‘20, this naturally means the team is not done delivering. This is only the beginning and much more is on the roadmap! If you are interested in understanding the roll-out of the new data prep platform check out this overview blog. I can also recommend checking out what Tim Bezold (Product Manager on the data platform) has to say about the features including his thoughts on what is on the roadmap in terms of new transformations.