How I Analyze a Neighborhood for Rental Property Cash Flow

In 2023 my portfolio collapsed. I'd scaled to 124 houses by buying the cheapest ones I could find and renovating them, and too many of them were on the wrong streets. The thirty I own today are the ones that were on the right ones.

This method is what came out of that. It's the report I run on every house before I look at the price, the lease or the loan. This page is the full version: the three phases, the eleven segments and their weights, the arithmetic, the geography rule, the crime and eviction rules, the red flags, and the nine questions I send to a property manager at the end. Every number here comes from three real reports I ran in October 2026, one for a property in each phase.

Why does trajectory beat level?

Because direction is what you own over a ten-year hold, and a snapshot is what you own for a month.

Two houses can sit at the same price, in the same ZIP, with the same school rating, the same crime level and the same flood score. One is in an area where incomes are rising, empty homes are being filled and crime is falling. The other is in an area where owners are leaving and renters are taking their place. The listing can't tell you which is which. The direction can.

Falling crime, rising owner-occupancy, declining poverty rates, and real income growth drive demand from tenants and from the next buyer. That demand is what produces rent growth and value growth over time.

Rising crime and falling owner-occupancy do the opposite, however good the renovation. So I grade the trend of each segment, not only its level. A weak area moving the right way can outscore a slightly better area standing still or declining.

I learned this lesson the expensive way myself back in 2023. Cheap is not the same as good value.

What are the three neighborhood phases?

Most investors grade neighborhoods A, B, C and D. That's fine, but personally I use phases instead, because a letter describes where an area is and a phase describes where it's going.

Why Phase 2? Because it's the best balance of risk and return (in my opinion).

Prices and rents are still within reach for residents (housing affordability), so the numbers can work for rental property investors from day one. And because the area is improving, there's a good chance of rising rents, better tenants and growing value over the hold. None of that is guaranteed.

Investing is a game of probabilities, and Phase 2 on a positive trajectory is where I think the odds are best. Phase 3 isn't a bad place to own; the risk is lower and so, usually, is the return. Phase 1 is where I lost money, and you couldn't pay me enough to own in that kind of neighborhood any more.

One thing to be clear about before we look at the math. This report grades the neighborhood, not the house. The price against comps, the lease against market, the appraisal, the tax and insurance, the city's rental rules and the yield all belong in a separate underwriting report.

That's the answer to a question I get asked a lot: why doesn't your neighborhood grade mention the cap rate? Because that's a grade for the house, not the neighborhood. The neighborhood is one half of my buy box. The house, the price and the rent are the other half, and you can test yourself on any property using my free rental property cash flow calculator.

What do I grade, and how is the score built?

Eleven segments. Each gets a level from 1 to 5, which is where it stands today, and a trend: improving is +1, steady is 0, worsening is −1. The segment score is (level + trend) × weight. The weights add up to 92, so the most a neighborhood can score is 6 × 92 = 552 points, and the published score is the total ÷ 552 × 100.

That might sound a little complicated, but the end result is that we have a score out of 100, properly weighted on the most important factors that drive neighborhood quality and risk, using the best available data for the most relevant area.

Crime carries the most weight because it's the thing a tenant lives with every day, and the thing that most decides who applies for your house. Also, in most of the C class markets where properties are cheaper, the crime rate is generally higher than the metro average.

In my experience, outside of personal events like a job or location change, tenants don't leave because the rent goes up. They leave because they don't like the house, the property manager or landlord, or the neighborhood. And as landlords, we want our good tenants to stay as long as possible. Tenant turnover is the one thing that will kill your rental property cash flow fastest.

Investment and development carries the least weight, because a project breaking ground tells you what might happen, while owners, incomes and the crime trend tell you what is happening. Also, I haven't met a tenant yet who rented from me because the local zoo opened a new exhibit, or a developer built an affordable housing block a few streets away. They have, however, rented from me because the house is close to where a major employer is based and expanding, like a hospital, manufacturer or other large business.

Here's the method run in full on a real report: a renovated 4-bed in North Town Fork Creek, Kansas City, 64130, graded in October 2026.

58.5 is a decent Phase 2 score, and it's a Phase 2 built on direction.

Seven of the eleven segments carry an upward trend. The levels on their own are ordinary: a 2 for sales, a 2 for income, a 2 for schools, a 2 for housing stock. What lifts the area into Phase 2 from Phase 1 is the trend.

Incomes are up 37.5% in real terms since the 2015 to 2019 Census period.

Poverty has fallen from 47.5% a decade ago to 17.6%.

The share of empty homes has halved, from 26.5% to 13.9%.

Crime within a short walk is down 29.7% this year against a 9.6% fall for the city.

The area was in trouble a few years back, and now it's coming out of it.

A few of the segments need a word on how I read them, because the obvious reading is often wrong.

Rents. The tract's Census median rent includes small apartments, so it understates what a large house lets for. I buy single family homes, not apartments, so I grade the rent level on live house listings by bedroom within about 1.5 miles, and use the tract median only for the trend. In North Town Fork Creek, 3-bed houses ask about $1,350 and 4-beds about $1,550, while the tract median is $838. And HUD's rent limit for a ZIP is a ceiling on what a voucher can pay. Remember though, HUD FMR rents are not your rent.

Income. I convert the earlier Census release to today's dollars before claiming a gain, and test whether the change is statistically clear. In Stayton Meadows (the Phase 3, B class house) household incomes are above the city median, but only rose 18.6% in nominal dollars and fell 3.4% after inflation, so it grades as flat but is still a higher income neighborhood. That's normal for higher income neighborhoods. The income growth happens when it comes from a lower-than-average starting point. The same trap sits in the jobs data: a rising share of jobs paying over $3,333 a month is partly just wages rising with inflation.

Vouchers. Voucher tenants are a safety net, not a rent upside. The housing authority's rent reasonableness test caps a voucher rent near market, so the value of Section 8 renters is in long tenancies, not higher rent. If you're being sold on a Section 8 tenant using FMR rents, you're going to be disappointed.

Development. Every project gets a status: seeking partners, broke ground, ideas only with no funding. The Truman Sports Complex is the standing example in this analysis. Both sports teams are set to leave by 2031 and the county's redevelopment framework has no money behind it yet, so I grade it steady, not improving. A university or a hospital next door is an anchor, not an improvement trend.

The public record. Beds, baths and square footage numbers online are often outdated, especially with renovated turnkey real estate where beds and baths might have been added recently. The North Town Fork Creek house was described as 4 bed, 2 bath and 1,250 square feet. The public record says 2 bed, 1 bath and 720. That's a matter for the underwriting report, inspection, and appraisal, not the neighborhood grade.

Which geography does each segment use? (hint: it's not ZIP code)

I mentioned this right at the start of this article, and it's where most remote investors get it wrong.

I use the tightest area around the house with reliable data. Why? Because one street or block varies considerably from the next. This is the most important rule in my method. Most of these C class markets where affordable housing lives are street by street. A ZIP doesn't show you that. Often, it can hide problems on a particular block as well as lead to missed opportunities.

A county eviction rate or a ZIP rent says almost nothing about one street. In one of my neighborhoods in Kansas City, more than half the evictions are from a single multifamily building. Take that out of the equation and the single family eviction rate is below the county level. Picking the right city gets you to a metro where the numbers can work, and the Kansas City ranking gets you to the neighborhoods I target there. Neither picks the street, and the street decides the outcome.

A census tract is a small area of about 1,200 to 8,000 people (averaging around 4,000 nationwide). By comparison, urban ZIP codes typically cover 20,000 to 50,000, and the biggest exceed 100,000. Kansas City's 64130 is about 20,000. The census tract is the right unit for incomes, owners and evictions because it's the smallest area the Census Bureau publishes those figures for.

Tract figures still carry wide margins, and I say so in every report. Where a change isn't statistically clear, the report says it isn't. Where the survey sample is tiny, I use the 2020 Census full count instead.

The jobs buffer has one exception. A commuter neighborhood's job base is typically wider. The Stayton Meadows house is 7.3 miles from downtown Kansas City, so a 5-mile circle excludes downtown, the Plaza and the main hospital cluster. Within 5 miles there are about 55,000 jobs, nearly flat over ten years. Within 10 miles there are 427,000, up 10.1%. Few jobs next to the house is not a weakness when 427,000 sit within a commute, so the rule is 10 miles when downtown is more than 5 miles away.

One consequence of the geography rule catches people out. Two houses in the same tract share nine of the eleven grades, because income, poverty, owners, vacancy, rents, vouchers, evictions, jobs and schools are all tract-level or wider. Crime and the red-flag check are the house's own, because they run on the quarter mile and the block. That's how two houses 1.2 miles apart can look the same by ZIP code, and vary wildly when you dig into the local data. It's also how a third address 0.3 miles away in the same tract carries a red flag. I've already published the proof, two houses on one street in the same ZIP, 1.2 miles apart, graded 58.5 (Phase 2) against 41.1 (Phase 1).

How do I grade crime, and why on direction?

On the quarter mile, on the trend, on all police reports, with a floor for genuinely low-crime areas. Here's each piece and the reason for it.

Why direction, not level. If I graded crime on its level, every Phase 2 neighborhood would fail by definition. Phase 2 neighborhoods are typically above the city average but improving. That's one of the reasons they're affordable, and it's what makes them Phase 2. I grade the direction, because direction is what you live with when you hold the property for 10 years, and falling crime is what draws the next tenant and the next buyer.

Why all reports, not violent crime. Violent counts at the quarter mile are small enough that one year swings them wildly. That's just noise. All reports, each counted once, give a count big enough to carry a trend. If total crime reports are fairly low, but there have been three burglaries on that block in the past year, that's a huge red flag.

Why the quarter mile. It's a 5 to 10 minute walk, the area a tenant actually lives with. The half mile and the city are context. Like I said, these markets are street by street. That's how prospective tenants and homebuyers look at it, so that's how I look at it as well.

The level. I compare the earliest full year of data, 2022, with this year's pace to date, adjusted for a full year. Down 25% or more is Level 5. Down 10 to 25% is Level 4. Within 10% either way is Level 3. Up 10 to 25% is Level 2. Up 25% or more is Level 1. Where the earliest year has fewer than 100 reports, I compare two-year averages instead, so one odd year can't decide the grade.

The trend. This year's count to date against the same period last year, measured against the city's own change. Up if the fall beats the city's by 5 points or more. Down if this year is up 10% or more. Steady otherwise. We're looking for neighborhoods that are getting consistently safer.

Here's the worked line for North Town Fork Creek. Within a quarter mile of the house: 115 reports in 2022, 137 in 2025, 111 in 2025 to 4 October, 78 in 2026 to 4 October. At that pace, this year's total will land around 96. From 2022 to the pace is 115 to 96, down 16.3%, which is Level 4. This year is down 29.7% against the city's 9.6%, so the trend is up. That's a real positive trend, and that's what I'm looking for.

The low-crime floor, and why it exists. The trajectory rule is built for areas where crime is a real issue. In a quiet safer area it breaks. Stayton Meadows has about 20 crime reports a year within the quarter mile. With such low numbers, one or two reports can swing the trend significantly, which is not representative of the area.

So the first check is volume. If the quarter mile's total crime report volume is below the city's average, the level is 5 regardless of trajectory.

Kansas City recorded 56,880 reports in 2025 across 314.9 square miles, which is 180.6 per square mile. A quarter-mile circle is 0.196 square miles, so the city average for that circle is 35.5 reports a year. Stayton Meadows runs at 20.7, so it's Level 5 outright. The trend is graded as usual: down 52% this year against the city's 9.6%. Every city needs its own benchmark using the same math.

What's reported but not graded. Homicides within the half mile, with distance and date. The house's own block since 2022. How the block compares with the streets around it. None of that is in the score, but it's the material the property manager questions are built on.

Where the rule can't run. Some cities publish a limited crime data feed. Toledo's public crime map for example carries six categories only, so there's no all-reports count and the low-crime floor can't be applied. On a Toledo report I completed for a client, I graded crime at Level 3, steady, on the level, about the city average. On the six categories alone the trajectory would have come out higher, because the fall is mostly vehicle crime. On the violent subset alone it would have come out lower, because shootings within the quarter mile went 4, 7, 12 from 2023 to 2025. Twelve shootings in a year that close to the house is not great news, especially as the count has risen three years straight.

How do I read evictions, and what are the red flags?

Evictions get a grade like everything else, and they also get a red-flag test. Then there are seven other flags, and a few of them override the whole score.

The eviction flag. Tract filings per renter household, against the same figure for the county. The test is this year's pace: complete months to date × 12 ÷ months, dropping a partial month, not a multi-year average. If the tract's pace is above twice the county's, the house is Buy with conditions. And the report always states the peak: the peak year's count and ratio, then the change from the peak to now.

North Town Fork Creek's tract filed 32 evictions from January to July 2026. That's an annual pace of 54.9 against 406 renter households, or 13.5%. Jackson County's pace is 7.31%.

So the tract is at 1.85 times the county. But the peak was 2024, at 93 filings and 2.71 times the county, and the pace is down 41% since then. That's the finding that matters: the tract was in trouble and is coming out of it.

Stayton Meadows files at 0.47 times the county, down 84% from its 2020 peak. The Toledo tract files at 0.77 times its city, which I suspect is down to a higher share of Section 8 tenancies, though the HUD voucher data for that tract is still pending.

Two things to keep straight. An eviction filing isn't an eviction. Nationally, Eviction Lab's 2000 to 2016 data shows about one filing a year per 17 renter households but one eviction per 40, so roughly 4 in 10 filings end with the tenant out.

A high filing rate still says more tenants fall behind. And a high rate is a screening instruction, not a verdict. In a tract filing at nearly twice the county, I verify income at three times the rent with pay stubs and bank statements, the same way I screen any applicant. I also want a manager who starts the process the day rent is late, because an eviction takes weeks even when it goes well, and the reserve has to cover those weeks.

The red flags. A few turn a good score into Avoid. The rest turn it into Buy with conditions, and the conditions are written into the report.

The 311, dangerous-building and vacant-parcel layers are what turn "the tract is recovering, the street is behind it" from an impression into a measurable statement.

On the North Town Fork Creek house, the tract is clearly improving. But the city referred a property across the street for clean-up in August 2026, and a condemned house sits at the demolition stage a fifth of a mile away. Neither is a red flag that overrides the score. Both go on the conditions list, with a drive-by before any client commits.

In Toledo, a vacant lot 75 feet from the duplex triggered the flag, and there were 41 vacant parcels within the quarter mile.

The 311 history (non-emergency service line) of the house itself reads the renovation story off the public record too. The Stayton Meadows house shows a work-without-a-permit complaint in March 2025 and a dangerous-structure ticket in January 2026, both resolved. That fits a derelict house renovated after a late-2025 sale, and it tells me which permits to ask for. Older houses carry their history in the record, if you look.

How sure is the score?

Every report carries a sensitivity table. A score is a judgment with its workings exposed, not a black box, and the honest answer to "how solid is a 58" is to show what would move it.

North Town Fork Creek is Phase 2 in every case I tested.

Stayton Meadows is Phase 3 by 3.9 points, and the technical report says it plainly: Phase 3 depends on two rule decisions, the low-crime floor and the 10-mile jobs buffer. Inside those rules it holds under every single-segment change.

Toledo sits 0.7 below the Phase 2 line, and the crime grade could move it either way. The recommendation to pass on that duplex rested on more than the score: the price, the costs and the vacant lot next door.

Where this year's data decides the crime trend, the report says what next month could do to it, and I re-run it when the next month of police data lands. The two Kansas City reports above get re-run in early November.

What can't the data tell you?

What happened on the street this year. Public data runs 12 to 24 months behind it, and a Census five-year estimate is an average over five years.

So every report ends with questions to a property manager who works in that neighborhood, and it has to be an independent one, not the seller's. How I vet a manager includes asking which streets they'd avoid, and this is where that question earns its keep.

Here are the nine, as I send them. Take them to your own manager.

1. How many units do you manage within a mile of this house?
2. How many days does it take to lease a house like this one here, and how many applicants per listing?
3. What share of applicants pass your screening? What minimum income-to-rent ratio do you enforce here, and how do you verify income?
4. What are vacancy, annual turnover and late-payment rates in your units here? How long does an eviction take in this county today?
5. What does this type of house actually lease for here, on the open market and on a voucher?
6. How does this block compare with the streets around it? Any you'd avoid?
7. What's happening at the specific addresses the 311 data flagged on this block?
8. Have yards or basements flooded on this block?
9. Would you buy on this block yourself? Score it 1 to 5 and give the reason.

Questions 3, 6 and 7 matter most. The screening answer tells you who the tenant pool really is. The street comparison is the one thing no dataset holds. And the 311 question turns a line in a city database into a person's account of the house across the road. Tenants buy the neighborhood and the street before they buy the house, and the manager leasing your competitors' houses knows that street better than any report. The data and the legwork work together.

What do the three phases look like side by side?

Here are the three October 2026 reports in one place.

• A duplex on the Ottawa edge of Toledo, 43607.
• A renovated 4-bed in North Town Fork Creek, Kansas City, 64130.
• A renovated 6-bed in Stayton Meadows, Kansas City, 64133.

The job of this section is to make the phases real places rather than definitions. The deal numbers stay out, because this is the method page. The full comparison with the yields is a separate article.

And the three scorecards, segment by segment:

Read the arrows before the numbers, because they tell the story.

Phase 2 is the column where most of the arrows point up. Seven of the eleven segments in North Town Fork Creek are improving, and the levels are mostly 2s and 3s. That's the definition of Phase 2 in numbers: ordinary today, moving the right way, and not yet in the price.

Phase 3 is the column where almost everything is steady. Eight of eleven segments in Stayton Meadows are →, and the levels are 4s and 5s. That's "the improvement is already in the price" shown as data. Income is 117% of the city's, poverty is 2.5%, 8 in 10 homes are owner-occupied, and houses sell at asking in three weeks. There's less to go wrong and less upside from improvement.

Phase 1 is the column where the trends are mixed and the one down arrow in the whole set appears, on Toledo's schools. Four segments are improving, six are steady, and the levels are mostly 2s. Real incomes are flat, poverty is 43%, most homes are rented, tenancies run 2 to 4 years, and shootings within the quarter mile rose through 2025. The tract has a university and a hospital within a mile, and it still grades Phase 1. Demand that holds a weak neighborhood steady is not the same as a neighborhood getting better. That one sentence is what separates the phase model from a demand model.

Notice what all three share: weak schools. The high schools rate 1, 1 and 2 out of 10. Schools are the one segment no phase fixes in these markets, which is why they carry 8 of 92 points and not more. If the school rating were your filter you'd have passed on all three, including the Phase 3 one. The reality is, in these markets, most of the schools rate this way. The only highly rated schools are in and around $350,000+ house neighborhoods, and that's not where we're finding any positive cash flow.

And notice the one the snapshot gets wrong. The Phase 2 tract has the highest eviction filing rate of the three, nearly twice the county. On that number alone you'd rank it last. On the trend, it's the tract coming right fastest, down 41% from its peak as poverty fell. Trajectory beats level, again.

What does a good neighborhood look like?

The bottom line

If I could give a new investor one habit, it would be this: grade the direction of the street, not the price of the house. As one of my mentors says, "you buy the dirt first, then the house."

I'd rather own in the area that's recovering than the one that's standing still, even when the recovering one has the scarier eviction number today and the standing-still one has a university next door.

I'd rather own a 58 score that's moving up than a 49 score that's moving sideways. And I'd rather own either of those for cash flow than a 74 where the improvement is already priced in, unless parking money safely is the goal, in which case the 74 is exactly right. There are deals I turn down, and the 49 with the 16% gross yield on paper was one of them.

Everything in this method is public data. You could pull all of it with a laptop and an afternoon. The work is knowing which geography to use for each segment, reading the trend instead of the latest number, and then sending the nine questions to someone who walks the street.

None of it makes a neighborhood safe. You can score a 58 and still get a bad tenant, a burst pipe and a slow eviction in the same year, and if you own rental property long enough you will. Investing isn't about certainties, it's a game of probabilities. This is how I shift them, and it's how I stopped buying the wrong house for a right-looking reason.