💡 FB Refi Marketing Project – Class 2 Recap
This week we showed how you can go beyond the obvious fields in ARMLS and pull “hidden” loan-type data from the public records feed. Once extracted, we joined it with Freddie Mac’s Weekly Survey to estimate likely interest rates — a quick way for loan officers to spot refinance opportunities before spending on ads.
What we actually did together:
-
Pulled ARMLS export files and isolated loan-type fields that aren’t obvious at first glance
-
Cleaned the file: removed duplicates, standardized dates, normalized dollar amounts
-
Appended Freddie Mac weekly rate data to estimate borrower interest rates
-
Created a “ready to upload” list for Facebook custom audiences
At the end we also covered:
-
How the script works step-by-step
-
Simple ways to speed up the process (batching, consistent file naming, pre-formatted CSVs)
-
Best practices for cleaning data before analysis
This is AAEL in action: no J.A.R.V.I.S. complex — just hands-on learning where you ask “how” and “why” while you work, and get immediate explanations you can apply on the next run.
📺 Watch the class video here:
🤝 View the ChatGPT Plus session we used: https://chatgpt.com/share/68d57202-3e54-8011-8104-3db207a677b1
📺 Follow along on my YouTube channel: youtube.com/@nhancedata
—
Robert Foreman
Doctoral Student, Educational Technology (DET) – Central Michigan University
📧 forem1r@cmich.edu | 📱 480-415-0783