NatureDepPrint

Inspiration

Every business depends on nature, but those dependencies are often hidden inside technical reports, academic research, and large datasets.

The paper Firm-Level Nature Dependence develops an international dataset measuring how listed firms depend on ecosystem services such as water supply, soil quality regulation, pollination, flood control, and climate regulation.

We wanted to turn that research into something students, educators, sustainability professionals, and business decision-makers could see, explore, and understand.

NatureDepPrint was inspired by one question:

How can we make firm-nature dependence visible?

What it does

NatureDepPrint is an interactive platform that helps users understand recorded company dependence on nature across 21 primary ecosystem-service dimensions.

Users can:

Learn how business activities connect to ecosystem services
Explore company-year records available in the supplied dataset
View a company's 21-service dependence fingerprint
Examine a company's history and compare company profiles
Review ranked ecosystem-service scores
Build a self-assessed fingerprint when a company is not in the dataset
Apply educational disaster-pressure scenarios and inspect their mathematical overlap with a recorded or self-assessed fingerprint
Read a concise research summary and search the ecosystem-service Data Dictionary

The platform distinguishes between nature dependence and nature-related risk.

A dependence score describes recorded reliance on ecosystem services. A nature-related risk realization also depends on factors that the scores do not measure, including ecosystem condition, location, disruption likelihood, regulation, and company adaptation capacity.

NatureDepPrint does not predict ecosystem damage, operational disruption, financial loss, or future events. It helps users identify where deeper assessment may be useful.

How we built it

We built NatureDepPrint as a static browser-based website that can be hosted on GitHub Pages.

The research paper covers 31,772 listed firms across 117 countries from 2010 to 2023. The cleaned source file currently used by the website is a subset containing:

222,564 firm-year records
26,195 firms identified by unique RICs
Data covering 2010 to 2023
21 primary ecosystem-service dimensions
4 additional cultural ecosystem-service fields, kept separate from the main fingerprint

Each company-year profile can be represented as a 21-dimensional vector:

$$ \mathbf{s} = [s_1, s_2, \ldots, s_{21}] $$

where each (sj) is the recorded company-level dependence score for one ecosystem service.

The website presents these fixed-order, fixed-scale values in several complementary views, including:

A vertical 21-bar fingerprint skyline in Learn
A radial profile in Explore
Ranked and source-order numerical views
Circular and spider-graph views for self-assessed Builder results

All primary-service visualizations use the same fixed 0-6 scale and service order so values remain comparable.

We also built a guided self-assessment inspired by the two considerations underlying ENCORE's activity-level dependency ratings: potential functionality loss and the difficulty or cost of adaptation. It is an educational questionnaire, not a reproduction of the paper's ENCORE, business-segment, and revenue-weighting methodology.

For each ecosystem service that a user marks as relevant, the Builder asks the user to score:

Potential functionality loss if the service became unavailable, unreliable, or substantially weakened
The difficulty or cost of adapting

The educational self-assessed dimension score is:

Self-assessed dimension score = Functionality-loss score + Adaptation score

Relevant dimensions receive integer scores from 2 to 6. Dimensions marked not relevant receive 0. These results are clearly labelled as self-assessed and are not official NatureDep scores.

The site also includes:

A guided Learn section
Search by company name and identifiers
Historical profile views
Company comparison tools
Educational disaster-pressure scenarios
A concise Research section
A searchable Data Dictionary

Challenges we ran into

Making a large dataset usable in the browser

The website dataset contains more than 220,000 firm-year records. Because the site is static, we preprocess the source data into browser-ready metadata, matrices, summaries, and history files so the application remains responsive without a backend.

Explaining 21 dimensions clearly

A 21-dimensional profile can quickly become difficult to read. We use complementary views for different tasks:

A vertical fingerprint for an accessible first reading
A radial profile for overall shape
Ranked scores and tables for numerical detail
Historical charts for changes across available years
A searchable Data Dictionary for definitions

Avoiding unsupported claims

The supplied dataset does not provide the complete information needed to establish:

Facility or supplier locations
Local ecosystem condition
Disruption probability
Company-specific adaptation capacity
Operational or financial losses

We therefore avoid presenting dependence scores or scenario overlaps as direct predictions of risk, damage, or loss.

Incomplete company coverage

The website contains many companies, but it does not contain every company. We addressed this with two pathways:

Explore an available company-year record
Build a separate self-assessed fingerprint

Separating research scores from self-assessed scores

The paper's company-level scores combine ENCORE activity ratings with business-segment revenue data. The Builder does not reproduce that methodology, so its results are labelled as user-provided, self-assessed values rather than official NatureDep scores.

Accomplishments that we're proud of

We are proud that NatureDepPrint:

Turns complex academic research into an interactive learning experience
Makes recorded firm-nature dependence visible
Preserves the fixed order of all 21 primary ecosystem-service dimensions
Supports both guided education and detailed record exploration
Allows users to inspect histories and compare company profiles
Provides a pathway for organizations outside the supplied dataset
Uses transparent, deterministic calculations
Keeps interpretation limits visible
Runs as a static website without a backend or external API

The feature we are most proud of is the Nature Fingerprint. It turns a complex vector of dependence scores into a visual profile whose individual services and overall shape can be inspected and compared.

What we learned

The most important lesson was that dependence is not the same as realized risk.

A company may depend strongly on an ecosystem service, but a risk realization also depends on ecosystem deterioration, the location and regulation of the relevant ecosystem, the likelihood of disruption, and the company's ability to adapt.

We also learned that one aggregate score is not enough. Two companies can have similar average recorded dependence while relying on very different combinations of ecosystem services. One profile may be concentrated in water-related services, while another may be concentrated in biological, soil, or habitat-related services. The complete fingerprint reveals those differences.

The paper therefore provides two complementary aggregate measures. NatureDepOverall summarizes revenue-weighted average dependence across relevant services, while NatureDepHigh is designed to capture strong dependencies without allowing high and low values to offset one another in an average. Neither measure replaces the full 21-service profile.

Finally, we learned that the same dataset can support both education and technical exploration when the interface keeps the underlying data, calculations, and limitations clear.

What's next for NatureDepPrint

A future extension could connect dependence profiles with additional data needed for fuller nature-risk assessment, such as:

Asset and supplier locations
Local ecosystem condition
Water stress and flood exposure
Nature-related incidents
Company adaptation measures
Operational and financial data

Those additions could help users study where recorded dependencies coincide with external conditions or company-specific exposure. They would still require carefully defined methods and should not be treated as automatic forecasts.

For now, NatureDepPrint focuses on the necessary first step: