Building automated valuation models (AVMs) that don't suck: A developer's guide

Alex Wilkinson

CEO of Houski

2026-01-18

Most Automated Valuation Models (AVMs) rely on incomplete datasets, simplistic algorithms, and black-box approaches that produce valuations professionals can't trust or explain to clients.

Comprehensive property datasets, accessible machine learning, and demand for transparency now give developers the tools to build AVMs that work. This guide shows you how. If you prefer a ready-made solution, Houski's AVM endpoint provides documented valuations with full accuracy transparency and Canadian market coverage.

Why most AVMs are inaccurate

Data Quality Issues

Incomplete coverage: Limited to Multiple Listing Service (MLS) listings or tax assessments, missing most of the market
Inconsistent measurements: Square footage calculated differently across jurisdictions
Missing renovations: Recent improvements that significantly impact value go unrecorded
Outdated information: Assessment data that's 2-5 years behind current property conditions
Location imprecision: Postal code-level data missing micro-location value drivers

Oversimplified modeling approaches

Simplistic regression models: Linear models trying to capture non-linear relationships
Insufficient location granularity: Treating entire neighborhoods as homogeneous
Inadequate feature engineering: Missing interaction effects between variables
One-size-fits-all modeling: Using the same approach for mansions and studio apartments
Ignorance of market segments: Luxury properties valued like standard homes
Temporal blindness: Failing to account for seasonal and cyclical patterns

Opacity without accountability

Hidden limitations: Pretending the model works everywhere
Missing context: Valuation without market trends or comparable listings
No feedback mechanisms: No way to correct obvious errors

Building a better AVM: The foundation

1. Comprehensive Data Foundation

Accuracy starts with comprehensive property data. The Canadian market now has detailed information on 19+ million properties through APIs like Houski. The block below is a live call against the real API, so the request code and the JSON response always reflect the current schema and current data:

API request

TypeScript code

const houski_data = async (): Promise<PropertiesResponse> => {

    // You must copy the PropertiesResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/properties');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('city', 'calgary');
    url.searchParams.set('country_abbreviation', 'ca');
    url.searchParams.set('province_abbreviation', 'ab');
    url.searchParams.set('select', 'interior_sq_m,bedroom,den,bathroom_full,bathroom_half,construction_year,land_area_sq_m,property_type,heating_type_first,foundation_type,roof_material,assessment_value,assessment_year,latitude,longitude');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: PropertiesResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 7.500000953674316,
  "data": [
    {
      "address": "31 Hawkside Park NW",
      "assessment_value": 648500,
      "assessment_year": 2026,
      "bathroom_full": 3,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 1988,
      "den": 1,
      "foundation_type": "Wood",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 126.34708404541016,
      "land_area_sq_m": 13583.0,
      "latitude": 51.12946319580078,
      "longitude": -114.17919921875,
      "property_id": "10000f97f5cb7b9f",
      "property_type": "Duplex",
      "roof_material": "Asphalt"
    },
    {
      "address": "6 1744 7 Street SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 2006,
      "den": 0,
      "foundation_type": "Wood",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 125.882568359375,
      "land_area_sq_m": 512.0,
      "latitude": 51.03611755371094,
      "longitude": -114.0792007446289,
      "property_id": "10004f7afe0c1946",
      "property_type": "House",
      "roof_material": "Asphalt"
    },
    {
      "address": "384 Copperpond Landng SE",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 2006,
      "den": 0,
      "foundation_type": "Wood",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 125.882568359375,
      "land_area_sq_m": 512.0,
      "latitude": 50.925743103027344,
      "longitude": -113.92975616455078,
      "property_id": "10007f9761f49940",
      "property_type": "House",
      "roof_material": "Asphalt"
    },
    {
      "address": "239 Dalhurst Way NW",
      "assessment_value": 1110000,
      "assessment_year": 2026,
      "bathroom_full": 2,
      "bathroom_half": 0,
      "bedroom": 3,
      "construction_year": 1971,
      "den": 1,
      "foundation_type": "Wood",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 108.46339416503906,
      "land_area_sq_m": 724.0999755859375,
      "latitude": 51.110740661621094,
      "longitude": -114.1513900756836,
      "property_id": "100086f6bc064d3f",
      "property_type": "House",
      "roof_material": "Asphalt"
    },
    {
      "address": "52 Cedargrove Way SW",
      "assessment_value": 662000,
      "assessment_year": 2026,
      "bathroom_full": 1,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 1984,
      "den": 0,
      "foundation_type": "Wood",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 136.4734344482422,
      "land_area_sq_m": 430.2999877929687,
      "latitude": 50.95145034790039,
      "longitude": -114.12571716308594,
      "property_id": "1000c277cd905d3b",
      "property_type": "House",
      "roof_material": "Asphalt"
    },
    {
      "address": "28 Sundown Gr SE",
      "assessment_value": 692500,
      "assessment_year": 2026,
      "bathroom_full": 3,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 1988,
      "den": 1,
      "foundation_type": "Wood",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 172.61241149902344,
      "land_area_sq_m": 436.0,
      "latitude": 50.899044036865234,
      "longitude": -114.04895782470705,
      "property_id": "1001109ab2aebbc0",
      "property_type": "House",
      "roof_material": "Asphalt"
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": true,
    "has_previous_page": false,
    "page_total": 113490
  },
  "price_quote": false,
  "result_total": 680936,
  "time_ms": 105,
  "ui_info": {
    "city": "Calgary",
    "city_id": "6ec95b53075d062c",
    "city_link": "ca/ab/calgary",
    "city_slug": "calgary",
    "country": "Canada",
    "country_abbreviation": "CA",
    "country_abbreviation_id": "9ace2b6431b7f1be",
    "country_abbreviation_link": "ca",
    "country_slug": "canada",
    "province": "Alberta",
    "province_abbreviation": "AB",
    "province_abbreviation_id": "aae1f05a0f89d2c7",
    "province_abbreviation_link": "ca/ab",
    "province_slug": "alberta"
  }
}

Essential data requirements:

Complete coverage: All properties, not just active listings
Rich attributes: 50+ standardized characteristics per property
Geographic precision: Exact coordinates for micro-location analysis
Construction details: Materials, age, and condition indicators
Market context: Recent listings, days on market, price adjustments

2. Sophisticated modeling approach

Modern AVMs need approaches that handle real estate's complexities:

Ensemble methods: Combining multiple model types for better performance
Geospatial modeling: Explicit handling of location effects
Market segmentation: Separate models for different property types and price points
Non-linear relationships: Capturing diminishing returns and threshold effects
Temporal dynamics: Accounting for seasonality and market cycles
Transfer learning: Leveraging patterns from data-rich areas to improve sparse regions

3. Transparent limitations

Honesty about what your model can't do builds credibility:

Confidence intervals: Communicating uncertainty appropriate to each prediction
Limitation disclosures: Being clear about where the model works best
Explainability: Providing insight into which factors drive the valuation
Comparable evidence: Showing similar properties that support the estimate
Feedback mechanisms: Allowing users to flag problematic valuations
Continuous validation: Regular backtesting against held-out data

Implementation: A developer's roadmap

If you'd rather skip the implementation, our predict endpoint returns the same shape of data through one API call.

Stage 1: Data acquisition and preparation

Build a property database using Houski's API. The block below is a live call against the real API showing the rich attribute set commonly pulled for training:

API request

TypeScript code

const houski_data = async (): Promise<PropertiesResponse> => {

    // You must copy the PropertiesResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/properties');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('city', 'calgary');
    url.searchParams.set('country_abbreviation', 'ca');
    url.searchParams.set('expand', 'listings');
    url.searchParams.set('province_abbreviation', 'ab');
    url.searchParams.set('select', 'latitude,longitude,property_type,construction_year,floor_above_ground,basement_type,basement_finish,roof_material_install_year,cooling_type_first,area_residential_list_price_per_sq_m,area_residential_rent_price_per_sq_m,area_commercial_list_price_per_sq_m,area_comparable_list_price_per_sq_m,area_comparable_property_tax_per_sq_m,area_comparable_assessment_value_per_sq_m,interior_sq_m,land_area_sq_m,land_frontage_m,land_depth_m,bedroom,den,bathroom_full,bathroom_half,heating_type_first,roof_material,foundation_type,fireplace,maintenance_fee,assessment_value,assessment_year,garage_parking_space_first,garage_type_first,demographic_income_median_pre_tax,demographic_household_size_1_person_percent,demographic_education_level_bachelors_degree_percent,demographic_dwellings_occupied_percent,demographic_transportation_car_truck_or_van_percent,demographic_municipal_population,demographic_age_median_of_the_population');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: PropertiesResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 21.94999122619629,
  "data": [
    {
      "address": "31 Hawkside Park NW",
      "area_commercial_list_price_per_sq_m": 5266,
      "area_comparable_assessment_value_per_sq_m": 4235,
      "area_comparable_list_price_per_sq_m": 4136,
      "area_comparable_property_tax_per_sq_m": 24,
      "area_residential_list_price_per_sq_m": 3702,
      "area_residential_rent_price_per_sq_m": 16,
      "assessment_value": 648500,
      "assessment_year": 2026,
      "basement_finish": "Finished",
      "basement_type": "Full",
      "bathroom_full": 3,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 1988,
      "cooling_type_first": "None",
      "demographic_age_median_of_the_population": 60.79999923706055,
      "demographic_dwellings_occupied_percent": 0.9751861095428468,
      "demographic_education_level_bachelors_degree_percent": 0.2956521809101105,
      "demographic_household_size_1_person_percent": 0.3846153914928436,
      "demographic_income_median_pre_tax": 58368,
      "demographic_municipal_population": 845,
      "demographic_transportation_car_truck_or_van_percent": 1.0,
      "den": 1,
      "fireplace": 1,
      "floor_above_ground": 1,
      "foundation_type": "Wood",
      "garage_parking_space_first": 2,
      "garage_type_first": "Attached",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 126.34708404541016,
      "land_area_sq_m": 13583.0,
      "land_depth_m": 187.2008056640625,
      "land_frontage_m": 72.5584487915039,
      "latitude": 51.12946319580078,
      "longitude": -114.17919921875,
      "maintenance_fee": 617,
      "property_id": "10000f97f5cb7b9f",
      "property_type": "Duplex",
      "roof_material": "Asphalt",
      "roof_material_install_year": 2023
    },
    {
      "address": "6 1744 7 Street SW",
      "area_commercial_list_price_per_sq_m": 5288,
      "area_comparable_assessment_value_per_sq_m": 2898,
      "area_comparable_list_price_per_sq_m": 7428,
      "area_comparable_property_tax_per_sq_m": 35,
      "area_residential_list_price_per_sq_m": 4492,
      "area_residential_rent_price_per_sq_m": 22,
      "basement_finish": "Finished",
      "basement_type": "Full",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 2006,
      "cooling_type_first": "None",
      "demographic_age_median_of_the_population": 34.79999923706055,
      "demographic_dwellings_occupied_percent": 0.8860103487968445,
      "demographic_education_level_bachelors_degree_percent": 0.4469696879386902,
      "demographic_household_size_1_person_percent": 0.5882353186607361,
      "demographic_income_median_pre_tax": 64000,
      "demographic_municipal_population": 825,
      "demographic_transportation_car_truck_or_van_percent": 0.65625,
      "den": 0,
      "fireplace": 0,
      "floor_above_ground": 2,
      "foundation_type": "Wood",
      "garage_parking_space_first": 0,
      "garage_type_first": "Attached",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 125.882568359375,
      "land_area_sq_m": 512.0,
      "land_depth_m": 36.34501266479492,
      "land_frontage_m": 14.087214469909668,
      "latitude": 51.03611755371094,
      "longitude": -114.0792007446289,
      "maintenance_fee": 386,
      "property_id": "10004f7afe0c1946",
      "property_type": "House",
      "roof_material": "Asphalt",
      "roof_material_install_year": 2023
    },
    {
      "address": "384 Copperpond Landng SE",
      "area_commercial_list_price_per_sq_m": 4943,
      "area_comparable_assessment_value_per_sq_m": 2898,
      "area_comparable_list_price_per_sq_m": 3296,
      "area_comparable_property_tax_per_sq_m": 20,
      "area_residential_list_price_per_sq_m": 3199,
      "area_residential_rent_price_per_sq_m": 18,
      "basement_finish": "Finished",
      "basement_type": "Full",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 2006,
      "cooling_type_first": "None",
      "demographic_age_median_of_the_population": 33.599998474121094,
      "demographic_dwellings_occupied_percent": 0.9604031443595886,
      "demographic_education_level_bachelors_degree_percent": 0.28032344579696655,
      "demographic_household_size_1_person_percent": 0.2631579041481018,
      "demographic_income_median_pre_tax": 68608,
      "demographic_municipal_population": 3440,
      "demographic_transportation_car_truck_or_van_percent": 0.934482753276825,
      "den": 0,
      "fireplace": 0,
      "floor_above_ground": 2,
      "foundation_type": "Wood",
      "garage_parking_space_first": 0,
      "garage_type_first": "Attached",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 125.882568359375,
      "land_area_sq_m": 512.0,
      "land_depth_m": 36.34501266479492,
      "land_frontage_m": 14.087214469909668,
      "latitude": 50.925743103027344,
      "longitude": -113.92975616455078,
      "maintenance_fee": 386,
      "property_id": "10007f9761f49940",
      "property_type": "House",
      "roof_material": "Asphalt",
      "roof_material_install_year": 2023
    },
    {
      "address": "239 Dalhurst Way NW",
      "area_commercial_list_price_per_sq_m": 5266,
      "area_comparable_assessment_value_per_sq_m": 3177,
      "area_comparable_list_price_per_sq_m": 6224,
      "area_comparable_property_tax_per_sq_m": 34,
      "area_residential_list_price_per_sq_m": 6182,
      "area_residential_rent_price_per_sq_m": 20,
      "assessment_value": 1110000,
      "assessment_year": 2026,
      "basement_finish": "Finished",
      "basement_type": "Full",
      "bathroom_full": 2,
      "bathroom_half": 0,
      "bedroom": 3,
      "construction_year": 1971,
      "cooling_type_first": "None",
      "demographic_age_median_of_the_population": 37.20000076293945,
      "demographic_dwellings_occupied_percent": 0.9367088675498962,
      "demographic_education_level_bachelors_degree_percent": 0.43396225571632385,
      "demographic_household_size_1_person_percent": 0.13793103396892548,
      "demographic_income_median_pre_tax": 56320,
      "demographic_transportation_car_truck_or_van_percent": 1.0,
      "den": 1,
      "fireplace": 1,
      "floor_above_ground": 1,
      "foundation_type": "Wood",
      "garage_parking_space_first": 0,
      "garage_type_first": "None",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 108.46339416503906,
      "land_area_sq_m": 724.0999755859375,
      "land_depth_m": 43.2224235534668,
      "land_frontage_m": 16.752878189086914,
      "latitude": 51.110740661621094,
      "longitude": -114.1513900756836,
      "maintenance_fee": 568,
      "property_id": "100086f6bc064d3f",
      "property_type": "House",
      "roof_material": "Asphalt",
      "roof_material_install_year": 2023
    },
    {
      "address": "52 Cedargrove Way SW",
      "area_commercial_list_price_per_sq_m": 5288,
      "area_comparable_assessment_value_per_sq_m": 3266,
      "area_comparable_list_price_per_sq_m": 4674,
      "area_comparable_property_tax_per_sq_m": 26,
      "area_residential_list_price_per_sq_m": 4789,
      "area_residential_rent_price_per_sq_m": 16,
      "assessment_value": 662000,
      "assessment_year": 2026,
      "basement_finish": "Finished",
      "basement_type": "Full",
      "bathroom_full": 1,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 1984,
      "cooling_type_first": "None",
      "demographic_age_median_of_the_population": 39.599998474121094,
      "demographic_dwellings_occupied_percent": 0.9748201370239258,
      "demographic_education_level_bachelors_degree_percent": 0.2098214328289032,
      "demographic_household_size_1_person_percent": 0.33128833770751953,
      "demographic_income_median_pre_tax": 54784,
      "demographic_municipal_population": 1810,
      "demographic_transportation_car_truck_or_van_percent": 0.8846153616905212,
      "den": 0,
      "fireplace": 1,
      "floor_above_ground": 2,
      "foundation_type": "Wood",
      "garage_parking_space_first": 0,
      "garage_type_first": "Attached",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 136.4734344482422,
      "land_area_sq_m": 430.2999877929687,
      "land_depth_m": 33.319271087646484,
      "land_frontage_m": 12.914446830749512,
      "latitude": 50.95145034790039,
      "listings": [
        {
          "expand_estimate_list_price": 574900,
          "expand_listing_date": "2022-07-26",
          "expand_listing_event": "Listed",
          "property_id": "1000c277cd905d3b"
        },
        {
          "expand_estimate_list_price": 555000,
          "expand_listing_date": "2022-08-04",
          "expand_listing_event": "Price decrease",
          "property_id": "1000c277cd905d3b"
        },
        {
          "expand_estimate_list_price": 539000,
          "expand_listing_date": "2022-08-30",
          "expand_listing_event": "Price decrease",
          "property_id": "1000c277cd905d3b"
        },
        {
          "expand_estimate_list_price": 525000,
          "expand_listing_date": "2022-09-08",
          "expand_listing_event": "Unlisted",
          "property_id": "1000c277cd905d3b"
        },
        {
          "expand_estimate_list_price": 539000,
          "expand_listing_date": "2022-09-15",
          "expand_listing_event": "Unlisted",
          "property_id": "1000c277cd905d3b"
        }
      ],
      "longitude": -114.12571716308594,
      "maintenance_fee": 370,
      "property_id": "1000c277cd905d3b",
      "property_type": "House",
      "roof_material": "Asphalt",
      "roof_material_install_year": 2019
    },
    {
      "address": "28 Sundown Gr SE",
      "area_commercial_list_price_per_sq_m": 4943,
      "area_comparable_assessment_value_per_sq_m": 4046,
      "area_comparable_list_price_per_sq_m": 5096,
      "area_comparable_property_tax_per_sq_m": 28,
      "area_residential_list_price_per_sq_m": 4838,
      "area_residential_rent_price_per_sq_m": 18,
      "assessment_value": 692500,
      "assessment_year": 2026,
      "basement_finish": "Finished",
      "basement_type": "Full",
      "bathroom_full": 3,
      "bathroom_half": 1,
      "bedroom": 3,
      "construction_year": 1988,
      "cooling_type_first": "Central",
      "demographic_age_median_of_the_population": 42.0,
      "demographic_dwellings_occupied_percent": 1.0,
      "demographic_education_level_bachelors_degree_percent": 0.2535211145877838,
      "demographic_household_size_1_person_percent": 0.10810811072587968,
      "demographic_income_median_pre_tax": 61440,
      "demographic_municipal_population": 545,
      "demographic_transportation_car_truck_or_van_percent": 0.945945918560028,
      "den": 1,
      "fireplace": 1,
      "floor_above_ground": 2,
      "foundation_type": "Wood",
      "garage_parking_space_first": 2,
      "garage_type_first": "Attached",
      "heating_type_first": "Gas furnace",
      "interior_sq_m": 172.61241149902344,
      "land_area_sq_m": 436.0,
      "land_depth_m": 33.53923034667969,
      "land_frontage_m": 12.999701499938965,
      "latitude": 50.899044036865234,
      "longitude": -114.04895782470705,
      "maintenance_fee": 386,
      "property_id": "1001109ab2aebbc0",
      "property_type": "House",
      "roof_material": "Asphalt",
      "roof_material_install_year": 2023
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": true,
    "has_previous_page": false,
    "page_total": 113490
  },
  "price_quote": false,
  "result_total": 680936,
  "time_ms": 234,
  "ui_info": {
    "city": "Calgary",
    "city_id": "6ec95b53075d062c",
    "city_link": "ca/ab/calgary",
    "city_slug": "calgary",
    "country": "Canada",
    "country_abbreviation": "CA",
    "country_abbreviation_id": "9ace2b6431b7f1be",
    "country_abbreviation_link": "ca",
    "country_slug": "canada",
    "province": "Alberta",
    "province_abbreviation": "AB",
    "province_abbreviation_id": "aae1f05a0f89d2c7",
    "province_abbreviation_link": "ca/ab",
    "province_slug": "alberta"
  }
}

Training data should include:

Core property attributes (from Houski's API): physical characteristics, location coordinates, building info, lot details
Listing history (Houski listings): list prices, days on market, price changes, listing types
Location context (Houski neighborhood data): schools, crime, walkability, transit, demographics

Stage 2: Feature engineering

Transform raw data into modeling-ready features:

Relative measurements: Houski area fields, size percentile within segment, age vs. surroundings, rooms vs. neighborhood average
Location features: distance to amenities, neighborhood price trends, school quality, crime relative to city, walkability, transit, and bicycle scores
Time-based features: days since recent comparable listing, market velocity, seasonal adjustment, neighborhood price trend, inventory over time
Interaction terms: size × neighborhood, age × renovation, bedrooms × property type, location × school district, condition × price point

Stage 3: Model development

Market segmentation: separate models by property type, price tier, region, and unusual cases
Model selection and ensemble: gradient boosting for main valuation, neural networks for interactions, geospatial models for location, linear baselines, random forests for importance
Hyperparameter optimization: temporal cross-validation, Bayesian tuning, mean absolute error (MAE) and median absolute percentage error (MAPE) rather than only root mean squared error (RMSE), validation across market conditions
Confidence modeling: predictive intervals, uncertainty quantification, property-specific scoring, market-condition impact

See our performance metrics on the AVM information page.

Stage 4: Deployment and transparency

API design: clear input/output docs, standardized error handling, rate limiting, model versioning
Explainability: feature importance per valuation, supporting comparables, market trends, key drivers
User feedback: error reporting, professional override workflows, continuous learning, performance tracking
Dashboard: confidence intervals, historical trends, similar properties, local indicators

Our predict endpoint lets you skip rolling your own AVM.

Houski's model accuracy: Transparent performance metrics

Accuracy metrics from our production models, measured on held-out test data the models never saw during training. Full breakdown on our AVM information page.

City-level models (highest accuracy)

City-specific models deliver the best performance by learning local market patterns:

City	R² Score	Median MAPE	Within 10%
Calgary, AB	0.95	4.6%	80.2%
Edmonton, AB	0.90	5.5%	71.8%
Vancouver, BC	0.93	5.7%	72.0%
Mississauga, ON	0.90	6.6%	66.5%
Vaughan, ON	0.91	6.6%	65.9%
Toronto, ON	0.85	8.0%	58.6%

Calgary's 4.6% median error means a $600,000 home is typically within $27,600. Roughly 80% of Calgary estimates fall within 10% of the actual value on held-out test data. Available through our predict endpoint for date-conditioned valuations across historical periods, or AVM endpoint for comprehensive reports with confidence scores.

Provincial models

When city models aren't available, provincial models provide strong regional accuracy:

Province	R² Score	Median MAPE	Within 10%
BC	0.83	7.6%	59.6%
MB	0.67	9.9%	50.5%
AB	0.69	10.0%	49.8%

National baseline (Canada)

For areas without sufficient local data:

Metric	Value
R² Score	0.70
Median MAPE	13.5%
Within 10%	40.0%

How this compares

Other Canadian AVM providers don't publicly disclose accuracy metrics. The best publicly reported off-market accuracy from major US providers sits around 7-8% median error.

Our city models outperform these benchmarks: Calgary at 4.6%, Edmonton at 5.5%, Vancouver at 5.7%. To our knowledge, these are the most accurate property valuations available in Canada. Numbers fluctuate as we refine models. Check the AVM information page for current metrics.

Understanding these metrics

Median MAPE is the typical error. 5% means half of estimates are within 5% of actual value. We use median rather than mean because outliers like $10M mansions shouldn't skew expectations for typical homes.

R² Score measures how much price variation the model explains. 0.95 means it accounts for 95% of price differences between properties.

Within 10%/20% answers: "How often will estimates be close enough for my use case?"

The hierarchical approach

The system automatically selects the most accurate available model:

City model (if trained for that city)
Provincial model (if no city model exists)
National model (fallback for sparse data areas)

We only deploy city models when they outperform provincial models, so you always get the best estimate possible.

See which model applies to any location on our AVM information page, and click into city models for full performance breakdowns.

Supercharging your AVM with geospatial context

Basic models use postal codes. Accurate valuations need granular location data. Houski's /geocoding and /map endpoints provide it.

1. Residential property context

Comparable properties in the immediate area are fundamental for any valuation. The block below is a live call returning residential context within 500 metres of a target coordinate:

API request

TypeScript code

const houski_data = async (): Promise<GeocodingResponse> => {

    // You must copy the GeocodingResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/geocoding');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('expand', 'listings');
    url.searchParams.set('latitude', '51.0447');
    url.searchParams.set('longitude', '-114.0719');
    url.searchParams.set('property_type_eq', 'House');
    url.searchParams.set('radius', '0.5');
    url.searchParams.set('select', 'latitude,longitude,interior_sq_m,bedroom,den,bathroom_full,bathroom_half,assessment_value,assessment_year');
    url.searchParams.set('shape', 'circle');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: GeocodingResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 4.440000057220459,
  "data": [
    {
      "address": "201 505 8 Avenue SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "den": 0,
      "interior_sq_m": 125.882568359375,
      "latitude": 51.045658111572266,
      "longitude": -114.0718231201172,
      "property_id": "328ddfdcd712fa26"
    },
    {
      "address": "305 505 8 Avenue SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "den": 0,
      "interior_sq_m": 125.882568359375,
      "latitude": 51.045658111572266,
      "longitude": -114.0718231201172,
      "property_id": "62614d31007508e9"
    },
    {
      "address": "200 505 8 Avenue SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "den": 0,
      "interior_sq_m": 125.882568359375,
      "latitude": 51.045658111572266,
      "longitude": -114.0718231201172,
      "property_id": "736257e5099784e0"
    },
    {
      "address": "500 505 8 Avenue SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "den": 0,
      "interior_sq_m": 125.882568359375,
      "latitude": 51.045658111572266,
      "longitude": -114.0718231201172,
      "property_id": "7ea72c951f47c018"
    },
    {
      "address": "310 505 8 Avenue SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "den": 0,
      "interior_sq_m": 125.882568359375,
      "latitude": 51.045658111572266,
      "longitude": -114.0718231201172,
      "property_id": "905b5b3e094e0bc0"
    },
    {
      "address": "515 505 8 Avenue SW",
      "bathroom_full": 2,
      "bathroom_half": 1,
      "bedroom": 3,
      "den": 0,
      "interior_sq_m": 125.882568359375,
      "latitude": 51.045658111572266,
      "longitude": -114.0718231201172,
      "property_id": "9412f76df2925649"
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": true,
    "has_previous_page": false,
    "page_total": 495
  },
  "price_quote": false,
  "result_total": 2968,
  "time_ms": 106
}

This returns attributes plus listing history:

Recent comparable listings
Price trends in the micro-neighborhood
Time on market for similar properties
Price adjustments during listings

2. Recent permit activity

Permits signal unrealized value or neighborhood transformation that traditional AVMs miss. The block below is a live call returning recent permit activity in the area:

API request

TypeScript code

const houski_data = async (): Promise<GeocodingResponse> => {

    // You must copy the GeocodingResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/geocoding');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('expand', 'permits');
    url.searchParams.set('expand_permit_application_date_gte', '2023-06-01');
    url.searchParams.set('filter_expand_match', 'all');
    url.searchParams.set('latitude', '51.0447');
    url.searchParams.set('longitude', '-114.0719');
    url.searchParams.set('property_type_in', 'House,Commercial');
    url.searchParams.set('radius', '0.5');
    url.searchParams.set('shape', 'circle');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: GeocodingResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 0.2800000011920929,
  "data": [
    {
      "address": "513 9 Avenue SW",
      "permits": [
        {
          "expand_permit_application_date": "2025-03-10",
          "expand_permit_content": "Status: Completed|Issued date: 2025-04-07T00:00:00.000|Completed date: 2025-07-04T00:00:00.000|Estimated project cost: 5000.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: N/A|contractor: N/A|estimated project cost: 5000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2025-04068",
          "expand_permit_type": "Building",
          "property_id": "e9baf46145362e25"
        },
        {
          "expand_permit_application_date": "2024-04-29",
          "expand_permit_content": "Status: Completed|Issued date: 2024-06-04T00:00:00.000|Completed date: 2024-07-04T00:00:00.000|Estimated project cost: 5000.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: EXPLORON|contractor: N/A|estimated project cost: 5000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2024-08042",
          "expand_permit_type": "Building",
          "property_id": "e9baf46145362e25"
        },
        {
          "expand_permit_application_date": "2026-06-18",
          "expand_permit_content": "Status: Completed|Issued date: 2026-06-24T00:00:00.000|Completed date: 2026-07-03T00:00:00.000|Estimated project cost: 5000.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: N/A|contractor: N/A|estimated project cost: 5000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2026-11251",
          "expand_permit_type": "Building",
          "property_id": "e9baf46145362e25"
        },
        {
          "expand_permit_application_date": "2024-01-19",
          "expand_permit_content": "Status: Released|Category: Signs - Discretionary Use - Third Party Signs|Description: TEMPORARY USE: SIGN - CLASS G (DIGITAL THIRD PARTY ADVERTISING SIGN - 3 YEARS)|Proposed use code: C2721|Proposed use description: SIGN - CLASS G|Permitted/discretionary: Discretionary|Land use district: CR20-C20/R20|Land use district description: Commercial Residential District|Concurrent LOC: N/A|Decision: Approval|Decision by: Development Authority|Decision date: 2024-03-05|Canceled/refused date: N/A|SDAB Number: N/A|SDAB hearing date: N/A|SDAB Decision: N/A|SDAB decision date: N/A",
          "expand_permit_id": "DP2024-00444",
          "expand_permit_type": "Development",
          "property_id": "e9baf46145362e25"
        }
      ],
      "property_id": "e9baf46145362e25"
    },
    {
      "address": "525 8 Avenue SW",
      "permits": [
        {
          "expand_permit_application_date": "2026-04-28",
          "expand_permit_content": "Status: Completed|Issued date: 2026-06-13T00:00:00.000|Completed date: 2026-06-29T00:00:00.000|Estimated project cost: 8716.75|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: TENT EVENT|contractor: N/A|estimated project cost: 8716.75|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2026-06987",
          "expand_permit_type": "Building",
          "property_id": "301a806c3b139f6"
        },
        {
          "expand_permit_application_date": "2023-06-08",
          "expand_permit_content": "Status: File Closed|Issued date: 2023-06-15T00:00:00.000|Completed date: 2023-11-16T00:00:00.000|Estimated project cost: 4500.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: SPECIAL EVENT RENTALS|contractor: N/A|estimated project cost: 4500.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2023-08836",
          "expand_permit_type": "Building",
          "property_id": "301a806c3b139f6"
        },
        {
          "expand_permit_application_date": "2023-08-28",
          "expand_permit_content": "Status: Completed|Issued date: 2023-08-30T00:00:00.000|Completed date: 2023-09-05T00:00:00.000|Estimated project cost: 4500.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: SPECIAL EVENT RENTALS|contractor: N/A|estimated project cost: 4500.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2023-14501",
          "expand_permit_type": "Building",
          "property_id": "301a806c3b139f6"
        },
        {
          "expand_permit_application_date": "2025-03-07",
          "expand_permit_content": "Status: Completed|Issued date: 2025-05-26T00:00:00.000|Completed date: 2025-06-30T00:00:00.000|Estimated project cost: 9100.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: N/A|contractor: N/A|estimated project cost: 9100.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2025-03971",
          "expand_permit_type": "Building",
          "property_id": "301a806c3b139f6"
        },
        {
          "expand_permit_application_date": "2024-06-12",
          "expand_permit_content": "Status: Expired|Issued date: 2024-07-10T00:00:00.000|Completed date: 2025-07-10T00:00:00.000|Estimated project cost: 5700.0|Permit class: Special Function Ten|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Temporary|Work class group: Unspecified|Work class mapped: N/A|applicant: SPECIAL EVENT RENTALS|contractor: N/A|estimated project cost: 5700.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2024-11636",
          "expand_permit_type": "Building",
          "property_id": "301a806c3b139f6"
        }
      ],
      "property_id": "301a806c3b139f6"
    },
    {
      "address": "404 9 Avenue SW",
      "permits": [
        {
          "expand_permit_application_date": "2023-06-23",
          "expand_permit_content": "Status: Completed|Issued date: 2023-06-27T00:00:00.000|Completed date: 2023-10-26T00:00:00.000|Estimated project cost: 72000.0|Permit class: Industrial|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Repair|Work class group: Improvement|Work class mapped: Existing|applicant: N/A|contractor: N/A|estimated project cost: 72000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2023-09977",
          "expand_permit_type": "Building",
          "property_id": "4b31244b21b9173a"
        },
        {
          "expand_permit_application_date": "2025-06-12",
          "expand_permit_content": "Status: Completed|Issued date: 2025-07-31T00:00:00.000|Completed date: 2026-04-07T00:00:00.000|Estimated project cost: 143000.0|Permit class: Industrial|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Repair|Work class group: Improvement|Work class mapped: Existing|applicant: C C D WESTERN|contractor: N/A|estimated project cost: 143000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2025-11279",
          "expand_permit_type": "Building",
          "property_id": "4b31244b21b9173a"
        },
        {
          "expand_permit_application_date": "2023-11-22",
          "expand_permit_content": "Status: Released|Category: Signs - Discretionary Use|Description: TEMPORARY USE: SIGN - CLASS E (DIGITAL MESSAGE SIGN) - 3 YEARS|Proposed use code: C2710|Proposed use description: SIGN - CLASS E|Permitted/discretionary: Discretionary|Land use district: CR20-C20/R20|Land use district description: Commercial Residential District|Concurrent LOC: N/A|Decision: Approval|Decision by: Development Authority|Decision date: 2023-12-04|Canceled/refused date: N/A|SDAB Number: N/A|SDAB hearing date: N/A|SDAB Decision: N/A|SDAB decision date: N/A",
          "expand_permit_id": "DP2023-08243",
          "expand_permit_type": "Development",
          "property_id": "4b31244b21b9173a"
        }
      ],
      "property_id": "4b31244b21b9173a"
    },
    {
      "address": "409 8 Avenue SW",
      "permits": [
        {
          "expand_permit_application_date": "2025-10-21",
          "expand_permit_content": "Status: Cancelled|Issued date: N/A|Completed date: 2025-11-04T00:00:00.000|Estimated project cost: N/A|Permit class: Industrial|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: New|Work class group: New|Work class mapped: New|applicant: N/A|contractor: N/A|estimated project cost: N/A|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2025-20604",
          "expand_permit_type": "Building",
          "property_id": "346e1324c2ac737c"
        }
      ],
      "property_id": "346e1324c2ac737c"
    },
    {
      "address": "4100 525 8 Avenue SW",
      "permits": [
        {
          "expand_permit_application_date": "2026-02-02",
          "expand_permit_content": "Status: Completed|Issued date: 2026-02-24T00:00:00.000|Completed date: 2026-04-21T00:00:00.000|Estimated project cost: 160000.0|Permit class: Commercial|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Alteration|Work class group: Improvement|Work class mapped: Existing|applicant: PRIME PROJECTS|contractor: PRIME PROJECTS|estimated project cost: 160000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2026-01516",
          "expand_permit_type": "Building",
          "property_id": "1457550bf2e11164"
        }
      ],
      "property_id": "1457550bf2e11164"
    },
    {
      "address": "4050 525 8 Avenue SW",
      "permits": [
        {
          "expand_permit_application_date": "2023-11-14",
          "expand_permit_content": "Status: Completed|Issued date: 2023-12-10T00:00:00.000|Completed date: 2024-04-16T00:00:00.000|Estimated project cost: 100000.0|Permit class: Commercial|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Alteration|Work class group: Improvement|Work class mapped: Existing|applicant: PRIME PROJECTS|contractor: PRIME PROJECTS|estimated project cost: 100000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2023-19625",
          "expand_permit_type": "Building",
          "property_id": "1ddd0c66b231abb7"
        },
        {
          "expand_permit_application_date": "2023-12-21",
          "expand_permit_content": "Status: Completed|Issued date: 2024-02-02T00:00:00.000|Completed date: 2024-04-16T00:00:00.000|Estimated project cost: 1271000.0|Permit class: Commercial|Permit type: Commercial / Multi Family Project|Permit type mapped: Building|Work class: Alteration|Work class group: Improvement|Work class mapped: Existing|applicant: PRIME PROJECTS|contractor: PRIME PROJECTS|estimated project cost: 1271000.0|Total square feet affected: N/A|Housing units affected: 0",
          "expand_permit_id": "BP2023-21843",
          "expand_permit_type": "Building",
          "property_id": "1ddd0c66b231abb7"
        }
      ],
      "property_id": "1ddd0c66b231abb7"
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": true,
    "has_previous_page": false,
    "page_total": 62
  },
  "price_quote": false,
  "result_total": 368,
  "time_ms": 257
}

Critical signals that most AVMs miss:

Recent renovations not yet in tax records
New commercial developments impacting residential values
Upcoming residential densification
Neighborhood building quality improvements

3. Amenity proximity

Nearby amenities impact values but are hard to quantify without specialized data. The block below is a live call returning commercial amenities within one kilometre:

API request

TypeScript code

const houski_data = async (): Promise<GeocodingResponse> => {

    // You must copy the GeocodingResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/geocoding');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('commercial_use_neq', 'Not applicable');
    url.searchParams.set('latitude', '51.0447');
    url.searchParams.set('longitude', '-114.0719');
    url.searchParams.set('radius', '1');
    url.searchParams.set('select', 'building_name,commercial_use,latitude,longitude');
    url.searchParams.set('shape', 'circle');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: GeocodingResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 0.9000000357627869,
  "data": [
    {
      "address": "345 10 Avenue SW",
      "building_name": "Craft Beer Market",
      "commercial_use": "Restaurant",
      "latitude": 51.04343032836914,
      "longitude": -114.0707550048828,
      "property_id": "423b33c34b270fa5"
    },
    {
      "address": "522 11 Avenue SW",
      "building_name": "West Coast Theatre Office Building",
      "commercial_use": "Culture center",
      "latitude": 51.04291915893555,
      "longitude": -114.07271575927734,
      "property_id": "6557ff585fafd50e"
    },
    {
      "address": "3400 421 7 Avenue SW",
      "commercial_use": "Office",
      "latitude": 51.04661560058594,
      "longitude": -114.07093048095705,
      "property_id": "154ec01740fe98bc"
    },
    {
      "address": "535 7 Avenue SW",
      "building_name": "City TV",
      "commercial_use": "Audio recording studio",
      "latitude": 51.04652786254883,
      "longitude": -114.0734634399414,
      "property_id": "eafa8d1508e0a410"
    },
    {
      "address": "617 8 Avenue SW",
      "building_name": "Globe Cinema",
      "commercial_use": "Cinema",
      "latitude": 51.04561233520508,
      "longitude": -114.07501220703124,
      "property_id": "7ebf16e342fe404c"
    },
    {
      "address": "522 6 Avenue SW",
      "building_name": "Local 522",
      "commercial_use": "Restaurant",
      "latitude": 51.048004150390625,
      "longitude": -114.07276153564452,
      "property_id": "4905d20dc4e0e6db"
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": true,
    "has_previous_page": false,
    "page_total": 13
  },
  "price_quote": false,
  "result_total": 75,
  "time_ms": 91
}

Quantifies key value factors:

Proximity to grocery stores, restaurants, retail
Access to medical facilities
Distance to employment centers
Desirable businesses (high-end vs. discount)

4. Neighborhood market statistics

Use the aggregate endpoint for neighborhood-level statistics. The block below is a live call returning a median estimate for a target community:

API request

TypeScript code

const houski_data = async (): Promise<AggregateResponse> => {

    // You must copy the AggregateResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/aggregate');
    url.searchParams.set('aggregation', 'median');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('city', 'calgary');
    url.searchParams.set('country_abbreviation', 'ca');
    url.searchParams.set('field', 'estimate_list_price');
    url.searchParams.set('province_abbreviation', 'ab');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: AggregateResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 1.0,
  "data": [
    {
      "aggregation": "median",
      "field": "estimate_list_price",
      "value": "627360.500"
    }
  ],
  "error": "",
  "price_quote": false,
  "time_ms": 56
}

Integrating geospatial context into your model

To incorporate location data into your AVM, you would orchestrate parallel calls to the properties endpoint for the subject, the geocoding endpoint for nearby residential context, permits, and amenities, and the aggregate endpoint for neighbourhood medians. The primary call against the subject property looks like this:

API request

TypeScript code

const houski_data = async (): Promise<PropertiesResponse> => {

    // You must copy the PropertiesResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/properties');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('expand', 'listings');
    url.searchParams.set('property_id_eq', 'bd9c6fb24c31c772');
    url.searchParams.set('select', 'address,latitude,longitude,community,city,province_abbreviation,country_abbreviation,estimate_list_price');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: PropertiesResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 0.2199999988079071,
  "data": [
    {
      "address": "302 610 17 Avenue SW",
      "city": "Calgary",
      "community": "Beltline",
      "country_abbreviation": "CA",
      "estimate_list_price": 304696,
      "latitude": 51.03819274902344,
      "listings": [
        {
          "expand_estimate_list_price": 325000,
          "expand_listing_date": "2023-08-21",
          "expand_listing_event": "Listed",
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_estimate_list_price": 315000,
          "expand_listing_date": "2023-10-12",
          "expand_listing_event": "Price decrease",
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_estimate_list_price": 299000,
          "expand_listing_date": "2023-11-11",
          "expand_listing_event": "Price decrease",
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_estimate_list_price": 299000,
          "expand_listing_date": "2023-12-18",
          "expand_listing_event": "Unlisted",
          "property_id": "bd9c6fb24c31c772"
        }
      ],
      "longitude": -114.07510375976562,
      "property_id": "bd9c6fb24c31c772",
      "province_abbreviation": "AB"
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": false,
    "has_previous_page": false,
    "page_total": 1
  },
  "price_quote": false,
  "result_total": 1,
  "time_ms": 40,
  "ui_info": {
    "address": "302 610 17 Avenue SW",
    "address_link": "ca/ab/calgary/beltline/302-610-17-avenue-sw",
    "address_slug": "302-610-17-avenue-sw",
    "city": "Calgary",
    "city_id": "6ec95b53075d062c",
    "city_link": "ca/ab/calgary",
    "city_slug": "calgary",
    "community": "Beltline",
    "community_id": "ecc51da246c7dd4a",
    "community_link": "ca/ab/calgary/beltline",
    "community_slug": "beltline",
    "country": "Canada",
    "country_abbreviation": "CA",
    "country_abbreviation_id": "9ace2b6431b7f1be",
    "country_abbreviation_link": "ca",
    "country_slug": "canada",
    "parent_address": "610 17 Avenue SW",
    "parent_property_id": "52a7d622eafe5319",
    "property_id": "bd9c6fb24c31c772",
    "province": "Alberta",
    "province_abbreviation": "AB",
    "province_abbreviation_id": "aae1f05a0f89d2c7",
    "province_abbreviation_link": "ca/ab",
    "province_slug": "alberta"
  }
}

Incorporating geospatial data captures factors traditional models miss:

Microneighborhood effects: Properties blocks apart can have very different values
Emerging trends: Increasing permit activity correlates with value increases
Amenity premiums: Quantify the exact premium for specific amenities
Development impact: Measure how new commercial development affects nearby values

Real-world example: A modular AVM architecture

A simplified example using Houski's property API. The architecture would loop through something like fetch property, classify segment, gather geospatial context, engineer features, run model, calculate confidence, find comparables. The primary underlying call against the property endpoint looks like this:

API request

TypeScript code

const houski_data = async (): Promise<PropertiesResponse> => {

    // You must copy the PropertiesResponse type declarations from the 
    // Houski API documentation to strongly type the response

    const url = new URL('https://api.houski.ca/properties');
    url.searchParams.set('api_key', 'YOUR_API_KEY');
    url.searchParams.set('expand', 'listings,assessments');
    url.searchParams.set('property_id_eq', 'bd9c6fb24c31c772');

    const response = await fetch(url);
    const data = await response.json();

    return data;
}

(async () => {
let data: PropertiesResponse = await houski_data();

// Log the response
console.log(data);
})();

API response

JSON

{
  "cache_hit": false,
  "cost_cents": 0.07999999821186066,
  "data": [
    {
      "address": "302 610 17 Avenue SW",
      "assessments": [
        {
          "expand_assessment_value": 301500,
          "expand_assessment_year": 2026,
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_assessment_value": 284000,
          "expand_assessment_year": 2014,
          "property_id": "bd9c6fb24c31c772"
        }
      ],
      "listings": [
        {
          "expand_estimate_list_price": 325000,
          "expand_listing_date": "2023-08-21",
          "expand_listing_event": "Listed",
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_estimate_list_price": 315000,
          "expand_listing_date": "2023-10-12",
          "expand_listing_event": "Price decrease",
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_estimate_list_price": 299000,
          "expand_listing_date": "2023-11-11",
          "expand_listing_event": "Price decrease",
          "property_id": "bd9c6fb24c31c772"
        },
        {
          "expand_estimate_list_price": 299000,
          "expand_listing_date": "2023-12-18",
          "expand_listing_event": "Unlisted",
          "property_id": "bd9c6fb24c31c772"
        }
      ],
      "property_id": "bd9c6fb24c31c772"
    }
  ],
  "error": "",
  "pagination": {
    "current_page": 1,
    "has_next_page": false,
    "has_previous_page": false,
    "page_total": 1
  },
  "price_quote": false,
  "result_total": 1,
  "time_ms": 64,
  "ui_info": {
    "address": "302 610 17 Avenue SW",
    "address_link": "ca/ab/calgary/beltline/302-610-17-avenue-sw",
    "address_slug": "302-610-17-avenue-sw",
    "city": "Calgary",
    "city_id": "6ec95b53075d062c",
    "city_link": "ca/ab/calgary",
    "city_slug": "calgary",
    "community": "Beltline",
    "community_id": "ecc51da246c7dd4a",
    "community_link": "ca/ab/calgary/beltline",
    "community_slug": "beltline",
    "country": "Canada",
    "country_abbreviation": "CA",
    "country_abbreviation_id": "9ace2b6431b7f1be",
    "country_abbreviation_link": "ca",
    "country_slug": "canada",
    "parent_address": "610 17 Avenue SW",
    "parent_property_id": "52a7d622eafe5319",
    "property_id": "bd9c6fb24c31c772",
    "province": "Alberta",
    "province_abbreviation": "AB",
    "province_abbreviation_id": "aae1f05a0f89d2c7",
    "province_abbreviation_link": "ca/ab",
    "province_slug": "alberta"
  }
}

Benefits:

Modularity: Components handle specific aspects
Flexibility: Models update independently
Transparency: Clear value derivation
Context: Information beyond the estimate
Confidence: Honest uncertainty communication

Beyond the algorithm: Making your AVM valuable

A great automated valuation is useful, not just accurate. Our predict endpoint returns the model's value at any date you pass in. The AVM endpoint delivers point-in-time valuations with confidence scores, model metrics, and comparables.

What the predict endpoint provides

Date-conditioned valuations: Ask the model what it would estimate for a property on any given date - strong for reconstructing historical trajectories, not for true forecasting (the model is a gradient-boosted tree and does not extrapolate past its training window)
Batch processing: Value multiple properties in a single request
Scenario modeling: Override property characteristics to test "what if" scenarios
Trend analysis: See how estimated values changed across past periods

What the AVM endpoint provides

Three valuations: List price, sale price, and monthly rent estimates
Confidence scores: Know how reliable each estimate is
Model metrics: Full accuracy statistics (R², MAPE, hit rates, error distributions)
Comparable properties: Up to 10 similar properties within 5km with list price and estimated sale date
Property overrides: Test how changes to bedrooms, size, etc. affect value
Generated reports: Shareable URL with full valuation details

Conclusion: The future belongs to transparent, accurate AVMs

The winners won't be those with black-box models, but those who combine accuracy with transparency. The next generation of AVMs will:

Explain key factors driving each valuation
Adapt to changing market conditions
Incorporate quantitative and qualitative information
Learn continuously from new market data and feedback

With Houski's API, developers have the data foundation to build automated valuation models that work without enterprise budgets or massive data science teams. To skip the build phase, our predict endpoint offers date-conditioned valuations across past and present, and the AVM endpoint provides comprehensive current valuations. Both use the same underlying models, with metrics on our AVM information page.

Start with our API documentation, or jump to the predict endpoint or AVM endpoint.

See more blog posts →

Account

Data products

Industries

Explore

Company

Account

Building automated valuation models (AVMs) that don't suck: A developer's guide

Why most AVMs are inaccurate

Data Quality Issues

Oversimplified modeling approaches

Opacity without accountability

Building a better AVM: The foundation

1. Comprehensive Data Foundation

2. Sophisticated modeling approach

3. Transparent limitations

Implementation: A developer's roadmap

Stage 1: Data acquisition and preparation

Stage 2: Feature engineering

Stage 3: Model development

Stage 4: Deployment and transparency

Houski's model accuracy: Transparent performance metrics

City-level models (highest accuracy)

Provincial models

National baseline (Canada)

How this compares

Understanding these metrics

The hierarchical approach

Supercharging your AVM with geospatial context

1. Residential property context

2. Recent permit activity

3. Amenity proximity

4. Neighborhood market statistics

Integrating geospatial context into your model

Real-world example: A modular AVM architecture

Beyond the algorithm: Making your AVM valuable

What the predict endpoint provides

What the AVM endpoint provides

Conclusion: The future belongs to transparent, accurate AVMs