Q&A with a Data Scientist
We recently sat down with Jason Monaco, Prophyts' resident data scientist, for a peek behind the proverbial curtain to look at what makes the Prophyts AI Prediction Engine tick. Monaco is part of a key team that's instrumental in the day-to-day back-end optimization and the implementation of new features in the Prophyts AI Prediction Engine.
1. Jason, what volume of data do we process on a weekly basis?
On a weekly basis, the Prophyts AI Prediction Engine crunches 2TB of intent data, which represents the content consumption of more than 2.8 million businesses across almost 4,000 Web sites.
2. That's a lot of data. What does that represent in number of user domains monitored?
At this point we monitor 5,537,632 unique domains, all registered in the United States, which is our current focus. These represent users who are consuming content, who are tied to a specific IP address, at the user's business.
3. Can you describe the data that Prophyts consumes?
Prophyts consumes content downloads, reviews and searches that are compiled and mapped using natural language processing (NLP) to determine semantic relevance and then scored.
4. How long does it take to process the data?
Our platform is built entirely upon cloud services and as such is able to scale appropriately to ensure we're able to deliver at any volume. On average the process takes approximately five to six hours processing time with our current cloud platform usage.
5. How does Prophyts take raw, unstructured data and turn that into a prediction?
The Prophyts AI Prediction Engine sifts through several years of historical data, billions of monthly page views and the reading activity of millions of potential buyers from thousands of Web sites. This process identifies patterns of intensity (measures the strength of data consumption), frequency (tracks the number of times an event repeated during a particular set of data) and duration across overlapping intervals. From this pattern the prediction emerges.
6. How does Prophyts determine the accuracy of a prediction?
The prediction values of these current companies are then referenced against past purchase data from similar companies to determine accuracy across the Prophyts Prediction Categories.
7. How frequently does the Prophyts AI Prediction Engine make predictions?
The Prophyts AI Prediction Engine processes predictions in weekly intervals ensuring our customers have up-to-date insight into the companies that are most likely to buy solutions in their product categories.
Learn how Prophyts can solve your revenue problems. Contact us today!
Posted by David Seymour on 05/14/2019 at 9:46 AM