Dynamic Real Estate Comparable Sales Engine

Pull similar comps within a milage( this needs to be able to modify each time base off the area, in philadelphia, we can give a very persize distance between the property because every property is different, but other city we will need to redefine it to make the comps accurate.
Find size similar to the target ( the property we are analysing) sqft needs to be 400 sqft +/- range (this can be adjust later)
Identify the condition of the comps ( the comparable property, what is the condition. New or not new 2 options
Compare to the target property ( 6 category of rehab needed to get to the ARV comps )
Identify similar property to be comparable property and find at least 2, maximum 5 of comparable property and then take average pricing per sqft, and use that to times the sqft of the target property and see the ARV of our specific property



Build a dynamic Comparable Sales Engine that:
Pulls geographically relevant comps


Filters by size similarity


Adjusts for condition (new vs non-new)


Scores rehab gap vs ARV comps


Selects 2–5 most relevant comps


Calculates ARV using averaged $/sqft


This must be configurable by city and neighborhood.

SYSTEM LOGIC
1 Geographic Radius Logic (Dynamic Radius System)
Problem:
Philadelphia rowhouses can require very tight comp radius (e.g., 0.1–0.3 miles),
but suburban markets may need 0.5–1.0+ miles.
Required Feature:
Radius must be:
Adjustable per city


Adjustable per neighborhood


Adjustable per property density


Override-able manually


Implementation Suggestion:
if city == "Philadelphia":
default_radius = 0.25 miles
elif density == "urban":
default_radius = 0.5 miles
elif density == "suburban":
default_radius = 1.0 miles
else:
default_radius = 1.5 miles
Allow:
Admin override input


Per-property radius override



2 Size Similarity Filter
Target Property Square Footage = T
Comp must satisfy:
Comp Sqft >= T - 400
Comp Sqft <= T + 400
Make ±400 adjustable in admin settings.
Example:
Target = 1,600 sqft
Comp range = 1,200–2,000 sqft
This prevents using 900 sqft homes to comp 2,000 sqft homes.

3 Condition Classification
Each comp must be tagged as:
NEW / FULLY RENOVATED


NOT NEW / AVERAGE CONDITION


Binary for now (expand later).
Condition detection methods:
Option A:
MLS keywords:
"fully renovated"


"new construction"


"brand new"


"gut rehab"


Option B:
Manual override field
Target property must also be tagged:
Current Condition (as-is)


After-Rehab Condition (planned level)



4. Rehab Gap Classification (6 Categories)
You want to measure the difference between:
Comp condition vs Target condition
Define 6 rehab categories:
Example categories:
Demo -> new construction
Major structural issues,
Total gut renovation, including front and back yard
Interior total renovation only
Cosmetic renovation only, no electric, plumbing, hvac needed(or minor fix)
Bathroom/Kitchen/floor upgrade only cosmetic pricing
Total rehab score calculated.
This does NOT change ARV.
It helps estimate rehab budget and risk.

5. Comparable Selection Logic
After filtering by:
Radius


Sqft range


Sold within X months (recommend 6–12 months)


Condition match (if ARV target is “fully renovated”, use renovated comps)


Then:
Sort by similarity score.
Similarity score example formula:
Similarity Score =
(Distance Weight * Distance Score) +
(Sqft Weight * Sqft Difference Score) +
(Condition Match Weight)
Select:
Minimum: 2 comps
Maximum: 5 comps
If >5 eligible:
Select top 5 most similar.
If <2:
Expand radius slightly (increment 0.1 miles until minimum 2 found).

6. ARV Calculation Logic
For each selected comp:
Price Per Sqft = Sold Price / Comp Sqft
Then:
Average PPSF = (Sum of PPSF of selected comps) / Number of comps
Then:
ARV = Average PPSF × Target Sqft
Output:
ARV value


Comp list used


Average PPSF


Median PPSF (optional, safer metric)



Full Flow Summary for Developer
Input Target Property:


Address


Sqft


Current condition


Planned condition


Determine Radius (dynamic)


Pull sold comps within:


Radius


Sold within last X months


Sqft ±400 range


Filter by condition type


Score similarity


Select 2–5 best comps


Calculate:


Average PPSF


ARV


Output structured report

Note to clarify logic

We will have to identify 2 facts, 1st is the size of the bathroom, etc., second is sqft interior, but there are going to be different types of comparible i want to add in the future, but the feature I am looking for is, I identify the type of deal first
, and when you looking for ARV, you have to have a selection of each type of deal, what they are looking for, and they are specifically looking for comps in those specific condition, so i want to make sure that this filter can be manually change easily


Determine Type of deal selection
Flips

3. Flips with total renovation, interior/exterior
Existing condition
Front is not maintained, obvious crack, structural issues, for example, wall falling apart, major grass, the door is boarded up
If inside imagine is pulled, the house is full of trash, old kitchen, no recess lighting ( led lighting) old panels, bathroom is old vanity and dirty tile and there are signs that all mechanical needs to be replaced, and floors are not leveled.

And the comps in the area is all new updated floor, kitchen, bathroom, everything inside and outside

Flips with cosmetic renovation

New construction
Multi-family (2-4 units)
Multi- family 4-9 units
Multi - family 10 units