At the heart of AI is the desire to gain more value from our data. To truly gain value from data, we need to understand it. With Data Stories, Akkio automatically surfaces the drivers of the target column you’ve selected, giving you greater insight into your data.
To get started, simply create a flow with any input data.
For this example, we’ll connect a stroke prediction dataset from Kaggle, as a CSV upload. Then, we’ll click “Add Step,” select the “Predict” step, and select “stroke” as the column to predict. Finally, we’ll hit “Create Predictive Model.”
We’ll then see the Data Stories for our target column of “stroke.”
Zooming in on this, we can see how various predictive attributes impact the likelihood of stroke. For instance, elderly, married, urban men are highly at-risk of stroke, while babies and children (especially female), are at very low risk of stroke.
You can connect and predict any data in the same way, and Akkio will automatically surface the drivers of your selected target column.
As another example, let's examine some data stories from the banking dataset. If we train a model to predict "age" we get the following data story:
This makes a lot of sense, you are more likely to be older if you are retired and married, and have completed a 4-year education. You are more likely to be younger if you are a single student who has only completed a high-school education.
We can also use data stories to derive insights about our business. Let's see the data story for the prediction target "subscription."
Here we find some really interesting insights into our data. For this particular bank promotion, older customers with cell-phones were more likely to signup, while younger blue-collar workers largely did not fit the offer. You can use this insight to better target your marketing campaign.
As you can see, Data Stories are a powerful machine learning analytics tool.