In this project, we are going to use Obviously AI to create Personas based off data. ObviouslyAI is a no-code data science tool that enables you to predict and analyze data with a click of a button.

With Personas, Obviously AI allows you to create profiles or, to use their terminology, personas, from your datasets. With this information, your team can better target personas and drive business based on these insights.

The interface of Obviously AI is pretty straightforward. On the left hand side you have your navigation column where you can Predict and Analyze. Obviously AI offers some datasets that you can play around with but when you're ready to upload your own dataset, you can click the plus sign next to My Datasets and upload a file here.

Today, we are going to look at an AT&T pre-existing data set provided through Obviously AI. Here is where you select the dataset you want to view and here is where you can select which table to look at if your dataset has more than one table. You have the option to pick a column to run predictions on or you can click this 'Ask in Natural Language' toggle, which allows you to type in your question.

Let's go ahead and look at Churn. In less than a minute, we have an overview of the drivers of Churn. Right now we are looking at the top 10 drivers and it looks like monthly charges are directly proportional to Churn. Let's select All here and see what happens. When we select All, we can see that tenure is inversely proportional. Then we have a whole bunch of drivers indifferent to Churn. You can click on these dots and get more information.

Let's click on tenure. We see here that month-to-month results in more churn and a 2 year contract vs. a 1 year contract significantly reduces churn. Clicking through these drivers you can start to get a lot of information on how to strategically package a service. Now that we've explored the drivers of churn, it's time to make some personas.

A good rule of thumb is to make some attributes the same across the board with one or two different attributes differing so you can accurately compare them. If two personas are completely different across attributes then it will be hard to get insights on the differentiating factors between them.

First, we'll make a persona for females, on month-to-month contracts with tech support Female - Tech Support Yes- MtM and save this persona. And now let's make one for the same except tech support will be no.  Now, let's set this up to export to our Google Drive. To do this, you would go to Settings → Integrations & API and hit Connect Google Drive

Now, I can export this to Google Drive and you will see the report has the included predictions. You can name this report. Something like Churn_MtM_April. Now you are ready to use your insights to make decisions as a team. Perhaps you dedicate time to following up with the people predicted likely to churn and then see what happens next month ー did it help? Or, you can offer tech support to Month to month users and see if this changes. As you get more data, you can create, update or delete your personas.

Now it's your turn. If you have access to some datasets, great. If not, you can use the data provided by Obviously AI, or you can use datasets provided by sites like Kaggle. Let us know what kind of insights you find ー did you learn something surprising? We look forward to hearing what you find!