Obviously AI allows you to make predictions on multiple entities at once with our Batch Predict feature.
Let's say you have trained the model to predict customer churn and used data from January 2020 to August 2020. Here you have a labeled dataset which consists of various features and the final target column signifying whether the customer has churned or not.
Next, you want to predict the customers who are going to churn in October 2020. Let's say you have 1,000 customers. You just need to upload the file containing the customer data for the month of October.
Obviously AI will then run the best algorithm on the data and will output a file containing the predictions for all the customers. The predictions will include a label signifying whether the customers have churned or not and will also include the customer churn probability.
To make the predictions on a batch follow the steps listed below:
Navigate to the Export Predictions tab on the Prediction Report.
In the Batch Predict tab click and select the file you want to upload. The best algorithm would then be used to make predictions on the batch data.
One the Predictions are ready you have the option to download the CSV file or export it to your Google Drive.
In case of the customer churn prediction problem there will be two new columns appended to the batch CSV file. One contains the label whether the customer has churned or not, other contains the churn probability.
The Obviously AI API
While our platform is built for non-technical users, our APIs are built for slightly more technical users looking to setup Predictions or Analytics for their team OR use Obviously AI's powerful features in their app.
With Obviously AI's APIs you can:
Export your prediction results to your internal dashboard.
Connect to external tools such as Zapier to automate tasks.
Give your users the ability to run predictions by integrating Auto-ML into your app.
Bring the power of machine learning to your phone or web app for your users.
Make predictions about your users in real time and setup responses. E.g. Dynamic pricing - show your users a price they're likely to pay.
If you'd like to discuss a use case, email us on [email protected].