A growing number of real estate companies are now using some form of AI-powered software. The adoption spans across a few use cases but none is more consequential for investors than deal screening.
The reason is straightforward. A single mid-sized US city can have thousands of active listings at any given moment. An investor, or a platform serving investors, can’t manually run cap rate and cash-on-cash return calculations on all of them just to find the handful that are worth a closer look. Something has to do the first pass.
That’s the job of a predictive score API. Instead of returning raw numbers that still require interpretation, it returns a score. The score is a single value that tells you whether a property is worth investigating further, calculated automatically from the underlying investment data.
We built Mashvisor’s Predictive Scores endpoint for exactly this workflow.
If you haven’t seen how it fits into the broader API, our overview of the Mashvisor API is a good starting point before diving into scoring specifically.
Key Takeaways
- A predictive score API converts underlying investment metrics (cap rate, cash-on-cash return, rental income, market trends) into a single score that ranks properties or neighborhoods by investment potential.
- Mashvisor’s Predictive Scores endpoint includes three distinct tools: Investment Likelihood (property-level), Mashmeter (neighborhood-level), and Property Recommender (preference-based matching).
- Scores are designed for filtering at scale, narrowing thousands of listings down to a shortlist before deeper analysis with the Investment Analysis endpoint.
- Common integrations include marketplace ranking, lead scoring for off-market deal sourcing, and automated alerts when a new listing scores above a threshold.
- Predictive scoring works best paired with the underlying investment data it’s built from, not as a replacement for it.
What Is a Predictive Score API?
A predictive score API takes the raw inputs behind a rental property’s financial performance (things like rental income, cap rate, cash-on-cash return, and local market trends), and condenses them into a single score or rating. The goal is to let a user, or an algorithm, quickly compare a large number of properties without evaluating each one’s full financial picture manually.
This is different from a basic investment analysis endpoint, which returns the underlying numbers themselves. A predictive score sits a layer above that. It’s the difference between an API telling you “this property has a 9.4% cap rate and 6.2% cash-on-cash return” versus “this property scores a 4.6 out of 5 for investment potential,” with the underlying numbers available as supporting detail.
For platforms with high listing volumes, that distinction matters a lot. Showing a user a ranked list based on a score is a far better experience than showing them a wall of raw percentages they’d need a finance background to interpret.
Inside Mashvisor’s Predictive Scores Endpoint
Mashvisor’s Predictive Scores endpoint is actually three related tools, each suited to a different part of the screening workflow.
Investment Likelihood
Investment Likelihood: a property-level score (returned as investment_likelihood_score) that evaluates a specific listing’s suitability as a rental investment, on a scale platforms typically display as a star rating.
This is the score you’d attach to an individual property card, similar to how a search platform might show a “deal score” next to a listing price.
Mashmeter
Mashmeter: a neighborhood-level score (returned as mashMeter) that evaluates the overall investment potential of an area as a percentage.
Where Investment Likelihood tells you about one property, Mashmeter tells you about the area it sits in. It’s the right tool for “where should I even be looking” decisions before you get to specific listings.
Property Recommender
Property Recommender: a matching tool that takes a user’s stated preferences for property type, bedroom count, budget range, and target home value, and returns properties that fit, ranked by how well they match.
This is the closest thing to a “for you” feed for rental investment listings, and it’s commonly paired with the other two scores so recommended properties also show their investment potential.
Used together, these three cover neighborhood-level screening (Mashmeter), property-level screening (Investment Likelihood), and personalized matching (Property Recommender). That’s most of the “where do I even start” problem that new investors and high-volume platforms both run into.
How the Scores Are Calculated
Mashvisor’s predictive scores are generated by weighing the same underlying factors a human analyst would look at: projected rental income, cap rate, cash-on-cash return, occupancy trends, and how a property or area compares to similar ones nearby. The models are continuously updated as new market data comes in, which is why the same property’s score can shift over time even if nothing about the property itself has changed.
Being direct about what scores are and aren’t matters here. A predictive score is a starting point for prioritization, not a substitute for full underwriting. A property scoring well on Investment Likelihood is worth pulling the full Investment Analysis data for. A property scoring poorly probably isn’t worth the time. That filtering step is the entire value.
How to Integrate Predictive Scores into a Deal-Screening Tool
The integration pattern is a funnel and each step narrows the candidate list using progressively more detailed data. All requests go to https://api.mashvisor.com/v1.1/client/ with your key passed in the x-api-key header.
Step 1: Identify neighborhoods worth screening
Start with the Get Top Neighborhoods endpoint to pull mashMeter scores across a city:
GET /v1.1/client/neighborhood/top/{state}/{city}
Each neighborhood object in the response includes a mashMeter field — an integer expressed as a percentage — alongside the neighborhood id you’ll need for the next step. Mashmeter is built on quantitative and qualitative data including average returns, occupancy, and overall market health for both short-term and long-term strategies. Filter to keep only neighborhoods above your threshold (a common starting point is mashMeter >= 50) and store the id values for Step 2. On a city like Nashville with 40+ neighborhoods, this typically reduces your working list to 8–12.
If you want a deeper profile of any neighborhood before proceeding, the Get Neighborhood Overview endpoint takes the neighborhood id and returns the full breakdown — Airbnb and traditional cap rate, rental income, occupancy, median price — with mashMeter included:
GET /v1.1/client/neighborhood/{id}/bar
Step 2: Score individual listings
For the neighborhoods that cleared Step 1, pull active listings:
GET /v1.1/client/city/listings?state={state}&city={city}&page=1&page_limit=50
Optional filters — beds, baths, min_price, max_price, property_type — let you narrow the set before scoring. Each property object in the response includes neighborhood_id (to cross-reference your Step 1 shortlist) and id (the identifier you pass to Investment Likelihood):
GET /v1.1/client/property/{id}/investment-likelihood?state={state}
The response returns investment_likelihood_score on a 0–5 scale. A threshold of >= 3.5 is a reasonable starting point for a deal-screening tool, but adjust based on how aggressively you want to filter. Properties below threshold get dropped; above threshold move to Step 3.
Step 3: Pull full investment analysis
For every property that cleared your score threshold, retrieve the detailed financial data:
GET /v1.1/client/property/{id}/investment?state={state}&payment_type=loan&down_payment={pct}&interest_rate={rate}&loan_type=1
The payment_type, down_payment, and interest_rate parameters let you model the actual financing scenario your users are underwriting. The response returns cap rate, cash-on-cash return, projected rental income, and occupancy estimates for both Airbnb and traditional strategies side by side. See our deeper walkthrough in the Investment Analysis API guide.
Step 4 (optional): Add personalization with Property Recommender
If you’re building a consumer-facing app rather than an internal acquisitions tool, Property Recommender matches listings against a user’s stated criteria — property type, bedroom count, target price range — and returns results ranked by fit:
GET /v1.1/client/property/recommend?state={state}&city={city}&property_type={type}&beds={n}&budget={max_price}
The recommended properties come back with enough identifiers to pass each one through Investment Likelihood in Step 2, so the personalized feed and the score-based filter work in series rather than as alternatives.
One implementation note: scores update daily, so a property’s investment_likelihood_score or a neighborhood’s mashMeter can shift between sessions. If you’re caching responses, keep TTLs at or under 24 hours to avoid surfacing stale signals.
Same caveat stands: the Investment Likelihood path (/investment-likelihood) and Property Recommender parameters should be confirmed against your API access before publishing, since that section of the docs wasn’t fully retrievable. Everything else is verified from the docs.
Real-World Use Cases for Predictive Scoring
Marketplace platforms use Investment Likelihood scores to add a “deal quality” indicator directly on listing cards, letting users sort by investment potential the same way they’d sort by price or bedroom count.
Off-market deal sourcing tools use Mashmeter at the neighborhood level to prioritize where to focus outreach. Contacting property owners is expensive and time-consuming enough that you want confidence the area justifies it. This pairs naturally with property ownership data for platforms doing skip tracing or owner outreach.
Automated alert systems use score thresholds to notify users when a newly listed property crosses a defined Investment Likelihood score, turning predictive scoring into a real-time deal flow feed rather than something a user has to check manually.
Portfolio acquisition teams at larger operators use Mashmeter to continuously re-rank target markets as conditions shift, since a neighborhood that scored well a year ago may score differently today. Markets move, and so do the signals behind the scores.
STR vs. LTR: How Predictive Scores Surface Both Sides
One of the more useful things about Mashvisor’s scoring system is that it doesn’t force a single strategy. Both Mashmeter and Investment Likelihood factor in performance data for Airbnb and traditional rentals separately, so a platform can surface properties that score well under one strategy but not the other.
This is where short-term rental regulations become relevant. A neighborhood can show a high Mashmeter score driven by Airbnb performance data, but if that city has since restricted short-term rentals in that zone, the score needs context. Before surfacing Airbnb return projections for a property, check the STR Regulations endpoint for that city — a neighborhood scoring well on Mashmeter for STR doesn’t tell you whether the city currently permits it:
GET /v1.1/client/airbnb-property/regulations?state={state}&city={city}
Pairing that call with your predictive scores keeps the system honest as local rules change, and avoids showing users Airbnb return projections for properties that can’t legally operate as short-term rentals.
Bottom Line
The hardest part of rental property investment usually isn’t the math on any one property. It’s figuring out which properties are even worth doing the math on. A predictive score API solves that first step, turning thousands of listings into a shortlist users can actually work with.
Whether you’re building a marketplace, an acquisitions pipeline, or an alert system, the pattern is the same: screen broad with Mashmeter, narrow with Investment Likelihood, then confirm with full investment data. If you’re ready to see how this fits into your platform, reach out about API access and we’ll walk through it together.
FAQ
What is a predictive score API in real estate?
A predictive score API takes underlying property and market data, including rental income, cap rate, and occupancy trends, and converts it into a single score or rating that indicates investment potential, allowing platforms to rank or filter large numbers of properties quickly without requiring users to interpret raw financial metrics.
What’s the difference between Investment Likelihood and Mashmeter?
Investment Likelihood is a property-level score that evaluates a specific listing’s potential as a rental investment. Mashmeter is a neighborhood-level score that evaluates the overall investment potential of an area. Investment Likelihood helps you choose a property; Mashmeter helps you choose where to look.
Can predictive scores replace full investment analysis?
No. Predictive scores are designed for filtering and prioritization. They help narrow a large list of properties down to the ones worth deeper analysis. For an actual investment decision, you’d still want the underlying cap rate, cash-on-cash return, and occupancy data from the Investment Analysis endpoint.
How often do predictive scores update?
Mashvisor’s predictive scores are recalculated as underlying market data updates, which happens on a daily basis. So a property’s score can change over time even without any change to the property itself.
What is Property Recommender used for?
Property Recommender matches properties to a user’s stated preferences, including property type, bedroom count, and target home value, returning ranked results based on how closely each property fits those criteria. It’s commonly integrated as a personalized recommendation feed within marketplace or search platforms.
How is Mashvisor’s scoring approach different from AirDNA or Rabbu?
AirDNA and Rabbu both offer market-level performance data that works well for short-term rental analysis. Mashvisor’s Predictive Scores endpoint is built around property and neighborhood-level scoring that spans both short-term and long-term rental performance, designed for filtering large volumes of listings across both strategies.