Portfolio analytics for risk, diversification, and capital allocation

The tenant you forgot you owned

A regional office landlord ran a routine check in early 2025 and found something uncomfortable: a single logistics tenant, spread across six separate buildings under slightly different legal entity names, accounted for 22 percent of total portfolio rent. On paper, no single lease looked scary. Aggregated, one bankruptcy could have wiped out nearly a quarter of income.

That is the entire point of portfolio analytics. Individual deals hide risk. Aggregation reveals it.

In this lesson you will build the logic for a dashboard over a 40-property portfolio that surfaces four things: geographic concentration, tenant concentration, weighted-average lease term, and where the next dollar of capital expenditurecapital expenditureCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète → (capexcapexCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète →, meaning money spent to improve or maintain a property) earns the best risk-adjusted return.

Start with one clean row per property

Before any chart, you need a tidy dataset. One row per property, consistent columns. This sounds trivial. It is where most real portfolios break, because data lives in 40 different Excel files with 40 different naming conventions.

The minimum useful schemaschemaA schema is the formal blueprint that defines how data is structured, named, typed, and related within a database, file, or message.Voir la définition complète →:

| Field | Example |

|---|---|

| property_id | TX-014 |

| market | Dallas |

| asset_type | Industrial |

| net_operating_income | 1,250,000 |

| top_tenant | Acme Logistics |

| tenant_parent | Acme Holdings |

| lease_expiry | 2029-06-30 |

| lease_annual_rent | 900,000 |

| market_value | 18,000,000 |

| debt_balance | 10,000,000 |

Note tenant_parent. That single column is what would have caught the 22 percent concentration above. Always mapmapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.Voir la définition complète → operating entities up to their ultimate parent, because that is who actually goes bankrupt.

Net operating income (NOI) is rental income minus operating expenses, before debt payments and taxes. It is the standard cash-flow measure for a property.

Concentration: the risk hiding in plain sight

Diversification is not a vibe. It is a measurable spread of income across markets and tenants. Two quick metrics do most of the work.

Percent of income by bucket. Group NOI by market, then by tenant_parent. Sort descending. If your top three markets are 70 percent of income, you are a regional bet, not a diversified portfolio. Neither is wrong, but you should know which one you are.

The Herfindahl-Hirschman Index (HHI). Borrowed from antitrust economics, HHI measures concentration by summing the squared shares. If one tenant is 100 percent of income, HHI is 10,000 (the maximum). If income is spread evenly across 10 tenants, HHI is 1,000. Lower is more diversified.

import pandas as pd

df = pd.read_csv("portfolio.csv")

# Tenant concentration by ultimate parent
tenant = (df.groupby("tenant_parent")["lease_annual_rent"]
            .sum()
            .sort_values(ascending=False))
shares = tenant / tenant.sum()
hhi = (shares.pow(2).sum()) * 10000

print(shares.head(5).mul(100).round(1))   # top 5 tenants, % of rent
print(f"Tenant HHI: {hhi:.0f}")

Run the same grouping on market for geographic HHI. Now you have two numbers a lender or investment committee immediately understands.

For a plain-language primer on the index itself, the U.S. Department of Justice keeps a short explainer on HHI.

Weighted-average lease term (WALT)

WALT is the average time remaining on your leases, weighted by how much rent each lease pays. It answers: how long is my income contracted for?

A five-year lease paying 1,000,000 dollars matters more than a ten-year lease paying 50,000 dollars, so you weight by rent, not by counting leases equally.

from datetime import date

df["lease_expiry"] = pd.to_datetime(df["lease_expiry"])
today = pd.Timestamp("2026-01-01")
df["years_left"] = (df["lease_expiry"] - today).dt.days / 365.25

walt = (df["years_left"] * df["lease_annual_rent"]).sum() / df["lease_annual_rent"].sum()
print(f"WALT: {walt:.1f} years")

A WALT of 7 years feels safe. But WALT hides timing. A portfolio can show a healthy 7-year WALT while 40 percent of rent expires in a single year. So pair WALT with a lease expiry schedule: rent expiring per calendar year, as a bar chart. That "wall of expiries" is often the single most important risk picture in the whole dashboard.

🎬 [VIDEO: "Understanding WALT and Lease Expiry Profiles" — youtube.com — a concise walkthrough of how commercial real estate analysts read lease maturity charts]

Turning risk into a picture

A dashboard is not 40 numbers. It is a few visuals that force a decision.

Four panels do the job:

1. Income by market (bar chart, sorted). Spot geographic bets instantly.

2. Income by tenant parent (bar chart, top 10). Spot the hidden 22 percent tenant.

3. Lease expiry schedule (rent expiring per year). Spot the wall.

4. Risk versus return scatter (each property a dot). We build this next.

Tools do not matter as much as discipline. Power BIBITechnologies and processes that turn raw data into actionable insights via reporting, dashboards and analysis, so teams can decide based on facts rather than intuition.Voir la définition complète →, Tableau, Looker Studio, or a simple Python notebook with matplotlib all work. The value is in the aggregation logic, not the software logo.

Where the next dollar of capexcapexCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète → should go

Now the interesting part. You have a limited capexcapexCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète → budget. Forty properties are competing for it. Which one earns the best risk-adjusted return?

Start with the raw return on a capexcapexCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète → project: incremental NOI divided by cost.

Suppose a 500,000 dollar amenity upgrade at a Dallas industrial asset is expected to lift NOI by 75,000 dollars per year. That is a 15 percent unlevered yield on cost. Attractive on its face.

But "risk-adjusted" means you discount that return by how uncertain it is. Two adjustments matter most in real estate:

Lease-up risk. If the incremental income depends on signing new tenants in a soft market, haircut it. A stabilized building with in-place demand is lower risk than a speculative renovation.

Concentration effect. Spending capexcapexCapital Expenditure (CapEx) is money spent to acquire, upgrade, or extend long-lived assets like equipment, property, or software that deliver value over multiple years.Voir la définition complète → on a building leased to your already-oversized tenant makes your portfolio riskier, even if the project return is high. Spending on a building in an underweight market improves diversification, which has value even at a slightly lower headline return.

A simple, transparent scoring approach:

df["capex_yield"] = df["incremental_noi"] / df["capex_cost"]

# Confidence factor: 1.0 = income already contracted, lower = speculative
# Diversification bonus: >1 if the asset is in an underweight market
df["risk_adj_score"] = (df["capex_yield"]
                        * df["confidence_factor"]
                        * df["diversification_bonus"])

ranked = df.sort_values("risk_adj_score", ascending=False)
print(ranked[["property_id", "market", "capex_yield", "risk_adj_score"]].head(10))

The point is not the exact formula. It is making the tradeoff explicit. A 15 percent yield with a 0.6 confidence factor (speculative) scores 9. A 12 percent yield with a 0.95 confidence factor (contracted) scores 11.4. The "lower return" project wins on a risk-adjusted basis, and you can defend that to a skeptical committee with one line of logic.

Keep the confidence and diversification inputs visible and editable. Reviewers should be able to challenge your assumptions, not just your arithmetic.

Reading the dashboard like an owner

A finished dashboard is only useful if it changes behavior. Three questions to ask every quarter:

Am I more concentrated than last quarter? Track tenant and market HHI over time. Concentration usually creeps up quietly as your best-performing assets grow in value.

When does my income wall arrive? Watch the largest single year of lease expiries. If it moves inside your refinancing window, that is a financing problem before it is a leasing problem.

Is capex flowing to the highest risk-adjusted score, or to the loudest asset manager? The scatter plot exists partly to remove politics from allocation.

One caution on data qualitydata qualityThe degree to which data is fit for purpose: accurate, complete, consistent, timely, valid and unique. Poor quality data undermines analytics, reporting and AI.Voir la définition complète →. A dashboard is only as honest as its inputs. If lease_expiry dates are stale or tenant_parent mappings are wrong, the whole picture lies with confidence. Reconcile the underlying rent roll (the master list of tenants, leases, and rents for a property) against the dashboard totals at least quarterly. If portfolio rent in the dashboard does not tie to the sum of the rent rolls, stop and fix that before trusting any chart.

A note on levered risk

Everything above used NOI and unlevered yields, which measure the property itself. Debt changes the risk profile sharply. A 60 percent loan-to-value (debt divided by value) asset and a 30 percent one can have identical NOI and wildly different risk. For a fuller portfolio view, layer in a debt maturity schedule alongside the lease expiry schedule. Two walls arriving in the same year, loan maturities and lease expiries, is the classic setup for distress.