Agent tools / Data provenance

Rental Strategy Agent

Turning recurring owner-revenue analysis into a repeatable, source-aware rental research workflow.

GnG Vacation · Sole developer

3

rental strategies: short, mid and long term

The problem

I began by reconciling property accounts and explaining monthly and year-over-year revenue changes to owners. Comparing rental strategies repeatedly led me to build a research agent that combines comparable listings, historical observations and financial calculations.

How it works

  1. 01

    Property

    Address, location and property attributes

  2. 02

    Comparables

    Airbnb and Zillow listing observations

  3. 03

    Calculation tools

    Rental assumptions and financial comparisons

  4. 04

    Report

    Sourced inputs, assumptions and missing values

What I built

  • Comparable-listing research based on geography and property attributes.
  • Listing-history storage with first-seen, last-seen and asking-price observations.
  • Financial calculation tools for comparing short-, mid- and long-term rental strategies.
  • Provenance and report checks that distinguish sourced figures, assumptions and unavailable inputs.

Engineering decisions

Preserve the history the listing page loses

Recording observations over time allows later analyses to use historical asking rents and listing lifecycle signals alongside currently available properties.

Keep estimates tied to evidence

A disappeared listing does not prove a signed lease, and its last asking price does not establish the final rent. The system uses those observations as signals in an estimate.

Give calculations their own tools

Financial tools and numeric provenance checks make the assumptions behind a recommendation inspectable, instead of relying on an untraceable generated number.