Introduction to Einstein Discovery

Einstein Discovery is an alternative of sophisticated Data Models which provide AI powered analytics to discover relevant patterns based on data provided by various stakeholders. Einstein Discovery is used to find insights and understand patterns across millions of rows of data in almost no time. Using Einstein Discovery, we can understand what happened, why it happened and what to do next.

Einstein Discovery works as a personal Data Scientist which analyzes large amount of data to learn patterns, understands correlations from it and thus helps in predicting best solutions.Based on this analysis, It provides efficient recommendations to stakeholders.

It provides answers to key business questions[1]:-

  1. What happened? Did something unusual happen?
  2. Why did it happen? What’s the diagnosis behind the facts?
  3. What will happen? Is there a trend, or is this just an isolated incident?
  4. What are some effective options for dealing with the situation?

Most importantly, it guides you in asking questions which may not be in your head while finding answers of above mentioned questions.

We can import data into Einstein discovery from almost all popular databases such as Hadoop, SQL, Salesforce, CSV files, Heroku, SAP etc.

Let's understand it through an example[1].

Suppose, you are VP of operations for a major automotive supplier, and one day you find that your margins are shrinking and you need to find out the reason for the same. Though you have a lot of data analysing which could predict the reason but it’s difficult to analyse that much amount of data manually using available tools, that too in a short span of time. Thus you need to use Einstein Discovery to carry out this task. You will need to upload your data to EInstein discovery and create stories to find insights and answers to all your questions.

Follow following steps to use Einstein Discovery for Shrinking Margin Problem:

  1. Download the dataset from (
  2. Launch Einstein Discovery and click on Datasets.
  3. Click on CSV, as our data format is CSV.
  4. Note: Einstein DISCOVERY is very smart, it can suggest some changes like spelling mistakes etc. If you feel suggestions correct then apply suggestions and click on apply otherwise leave them as it is.
  5. Note: You can also change type of Field Names.
  6. Once the dataset is created, click on datasets and then click on your uploaded dataset. A story setup screen would open.
  7. Choose the action variables like in our case it is MAXIMIZE THE VARIABLE MARGIN. Click on create story.
  8. After the story is created, a visualization of your dataset would be visible. This would tell about different insights about your dataset based on questions like What Happened ?, Why it Happened ? etc.
  9. Add Insights related to your solution on story.
  10. Export and share your story in Powerpoint, Docs etc.

Einstein Discovery automatically provides an explanation to any deviation in pattern of data and makes recommendation based on the data provided to it, thus eradicating the need of hiring a data scientist for your company. It is like having your own personal data scientist responsible for heavy lifting in order to speed up your work.

REFERENCES /wave_smart_data_discovery_your_data_scientist