How Scores Work

The six bars currently measure Safety, Schools, Activities, Food & Nightlife, Transportation, and Affordability. Dining and Nightlife are separate components of the Food & Nightlife bar. Housing and Cost of Living are not separate scored categories.

Category weights are relative: Safety 25, Affordability 20, Schools 15, Transportation 10, Activities 10, and Food & Nightlife 5. With all six available, their shares are approximately 29.41%, 23.53%, 17.65%, 11.76%, 11.76%, and 5.88%. We divide the weighted total by the sum of available weights. Missing categories are excluded, so profiles with different coverage may not be directly comparable.

The calculation stores scores on a five-point scale and multiplies them by two for display out of ten. Several current category formulas have a minimum of 1/5 (2/10). A displayed score is our calculated index, not an official government rating or a national percentile.

Overall = objective score × (1 − review influence) + adjusted review score × review influence. When no eligible reviews exist, the objective score supplies the whole rating. Reviews alone do not produce an official overall score.

The calculation uses violent, property, and other crime grades supplied through the Zyla crime API. We currently weight them 60%, 30%, and 10%. The displayed incident rates provide context; they are not directly used in this formula.

Each grade converts to a displayed score: A+ 9.0, A 8.6, A− 8.0, B+ 7.6, B 7.0, B− 6.4, C+ 5.8, C 5.2, C− 4.6, D+ 4.0, D 3.4, D− 2.8, F 2.0. Safety = 0.60 × violent score + 0.30 × property score + 0.10 × other score. All three grades are required.

This is a crime indicator. It does not measure every aspect of personal safety or predict what will happen at a particular address. Check the source geography and collection date on the area page.

SchoolDigger ratings are the main source. The school-selection pipeline also uses available school-provider evidence and identifies its source on the profile. The score averages valid selected school ratings on a 0–10 scale; it does not weight them by enrollment.

Searches can expand from 5 to 15 to 25 miles to find public-school coverage. Selected schools and their distances appear in the Schools panel. A nearby school is not a guarantee that a home is assigned to it. Confirm attendance eligibility with the district.

Provider ratings are narrower than a complete assessment of teaching, student support, or a child’s needs. No valid school ratings means N/A.

Google Maps search samples, accessed through Zyla, supply venue types, coordinates, ratings, and review counts. Activities include outdoor recreation, parks, cultural venues, and entertainment. Venues are classified from provider types, and duplicates and closed listings are removed.

The current formula is 20% quality, 40% supported variety, and 40% breadth. Breadth = min(weighted venue count ÷ 40, 1). Each variety family earns progressively more credit up to three weighted venues; family credits are averaged. Variety and breadth convert to 0–10 as 2 + 8 × component.

Quality uses review-adjusted Google ratings. For each rated venue: adjusted stars = (review count × stars + 20 × 3.5) ÷ (review count + 20). Average adjusted stars using geographic weights, then multiply by two. Review count adjusts confidence in each venue; it does not let a popular venue receive unlimited area weight.

For Jacksonville, Florida, Activities and Dining currently blend 90% of the local result with a 10% citywide reference derived from an excitement score of 14.3/100 (3.144/10). That is a citywide proxy, not ZIP-specific evidence. Other cities receive no such adjustment.

Food & Nightlife = 50% Dining + 50% Nightlife. Both components must have sufficient evidence; if one is unavailable, the combined score is unavailable. A prior saved score may remain visible during a refresh and is identified as older evidence.

Dining covers qualifying restaurants and cafés. It uses 20% quality, 40% supported variety, and 40% breadth, with a breadth target of 45 weighted venues. Quality and variety follow the rules described under Activities.

Nightlife currently uses 10% quality, 20% variety, and 70% evidence-weighted breadth. Its breadth target is 30. Dedicated club, music, and comedy venues count at full strength; other bar-first venues count at half strength; food-first venues with secondary bar types count at 0.15 strength. Geographic discounts also apply.

Nightlife variety covers clubs, music, comedy, breweries, pubs, and cocktails. Each family earns credit up to three evidence-weighted venues, then family credits are averaged. Variety and breadth convert to 0–10 as 2 + 8 × component. Quality is twice the geographically weighted average adjusted Google stars.

When fewer than three geographically weighted dedicated nightlife venues are found, Nightlife is capped at 6/10. This cap applies after the weighted formula. Hours, event frequency, attendance, and crowd activity are not measured by the current formula; the score describes sampled venue options.

The source is Census ACS mean travel time for workers who commute. Displayed score = clamp(14 − commute minutes ÷ 5, 2, 10). A 20-minute commute scores 10; 28 minutes scores 8.4; 40 minutes scores 6; 60 minutes scores 2.

Public-transit use, walking to work, and working from home are shown as context but do not currently change this score. Service frequency, traffic reliability, walkability, and access to jobs are not measured. Missing commute evidence means N/A.

Census ACS supplies median gross rent, renter household income, median home value, owner household income, and poverty rate. The score weights renter affordability 45%, owner affordability 45%, and poverty rate 10%. Available components are renormalized when inputs are missing.

Renter ratio = annual median gross rent ÷ median renter household income. Ratios of 20%, 35%, and 65% map to 10, 6, and 2 points. Owner ratio = median home value ÷ median owner household income. Ratios of 3, 5, and 9 map to 10, 6, and 2. Poverty rates of 5%, 15%, and 30% map to 10, 6, and 2. Scores interpolate linearly between these anchors and stay between 2 and 10.

This measures relative housing affordability, not a complete cost-of-living budget. It does not directly include mortgage interest, taxes, insurance, groceries, transportation costs, or household-specific spending. Geography may be broader than the ZIP where ZIP evidence is unavailable; the profile identifies the scope.

Census ZIP Code Tabulation Areas approximate postal ZIPs. Places inside that boundary receive full geographic weight. If fewer than three qualifying places are found in the category’s in-ZIP sample, places outside the boundary can contribute up to five miles away. The current trigger uses the category total, not separate Dining and Nightlife counts.

Outside weight = 0.5 × (1 − miles outside the boundary ÷ 5), with zero weight at five miles and beyond. For example, a venue one mile outside receives 0.4 weight; three miles outside receives 0.2. Distance is straight-line distance from the boundary, not driving time.

The search returns limited samples, not a complete inventory. Each search can request up to 40 results before filtering and deduplication. A venue absent from the sample is not proof that it does not exist. Without a verified boundary, a ZIP-only venue score cannot be calculated.

The current calculation includes eligible approved resident reviews and active starter perspectives, including AI-generated or uploaded starter entries. Public-discussion excerpts are excluded. Starter perspectives are labeled separately in the app and replaced as resident reviews are approved.

Influence anchors are: zero entries 0%; three 2%; seven 4%; fifteen 7%; twenty-five 10%; fifty 12.5%; seventy-five or more 15%. Between anchors, the weight increases smoothly using t² × (3 − 2t), where t is progress between the two review-count anchors.

The review average is adjusted toward the objective score using a 25-entry prior: adjusted review score = (count × review average + 25 × objective score) ÷ (count + 25). Approved resident votes are deduplicated by reviewer and area. Venue Google reviews are separate and affect venue quality instead.

The ranking compares ready profiles with positive objective and overall scores in the same state. It is recalculated when the area page is requested. The fraction at or above the area’s score includes ties and rounds up to a whole percent. At least 20 qualifying profiles are required.

Top 5% is shown for a top fraction of 5% or less. Fractions above 5% through 15% display “Around the top 10%.” Fractions above 15% through 20% display “Top 20%.” Lower-ranked profiles receive no top-percent badge. These are rated areas in our database, not all homes or every ZIP in the state.

Share cards use the saved area scores. Highlights show selected categories; the overall-score toggle changes presentation, not the underlying calculation. Follow the profile link to see the complete rating and its data coverage.

N/A means the score could not be supported by the available evidence. Venue scoring requires at least three qualifying places, at least three with valid ratings and positive review counts, and at least ten geographically weighted reviews. These gates apply to Dining and Nightlife separately as well as Activities.

Searches and refreshes check saved data and scoring versions, then request missing or outdated evidence. Place samples are considered expired after 30 days. Provider request limits can delay retries by approximately five to seven minutes; other failures have different retry rules. There is no guaranteed update time.

An older usable score may be retained while new evidence is incomplete. Source dates and geography describe the evidence used. Demographic attributes are descriptive and are not directly included as scored categories; income and poverty are used in the Affordability formula described above.

These methods describe the current implementation as of October 11, 2026. The weights and thresholds are product choices, not independently validated guarantees. Older saved profiles may use earlier versions until refreshed.