When the Data Lies and the Project Works Anyway
How to navigate the messiness of sustainability modeling without losing the plot.
Three datasets. Three different reputable data providers. Same company.
Three completely different sustainability ratings.
At first, I thought I’d made a mistake. But digging deeper, I found the real problem: the data itself.
In sustainability modeling, we work with imperfect, inconsistent, and often misleading inputs.
That doesn’t mean we stop modeling. It means we start listening harder — to the feedback loops, the ground truth, the pieces that don’t fit.
Because sometimes, the project works even when the codes and spreadsheets says it shouldn’t.
Sustainability Data Is a Mess — and Always Has Been
It’s tempting to believe that data in our field is clean, consistent, and comparable. But the truth?
Most sustainability data is:
Voluntary: Companies choose what (and whether) to report
Inconsistent: One firm calls it “waste,” another “by-product”
Self-reported: With all the incentives that come with that
Skewed toward what’s easy to measure, not what actually matters
And yet — we treat it like gospel. We plug it into models, publish results, build dashboards.
There is a difference between having numbers — and having signal.
And sustainability modeling should be in the business of finding signal.
The Myth of Clean Data
Here are two myths I still see — even among experienced analysts:
Myth 1: If it’s from a big provider, it must be accurate.
Reality: Different top-tier ESG data vendors routinely disagree — not just on ratings, but on raw inputs.
Myth 2: Clean data leads to strong conclusions.
Reality: “Clean” often just means standardized — and that can erase context. A clean emissions number tells you nothing about local impact, community risk, or adaptation strategy.
Clean data can still hide dirty truths — and messy data can sometimes reveal them.
This doesn’t mean we throw the numbers out. It means we treat them with the caution they deserve.
What We Do Instead: Modeling With Context
At Wangari, we take a different approach — one rooted in humility, context, and noise-tolerant design.
Here’s what that looks like:
We merge messy sources: local government reports, development bank disclosures, stakeholder interviews, NGO publications — even PDFs scraped from construction projects
We use proxies: If we don’t have a company’s water use, but we know the tech it uses and local supply stress, we infer likely impact
We build feedback loops: When a client corrects or improves the data, we revise the model — sometimes automatically, sometimes with human review
Missing data? We flag it, simulate ranges, or interpolate based on sector patterns.
Contradictory metrics? We weight them based on provenance and reliability.
Unstructured input? We parse it, index it, and link it back to meaningful indicators.
We don’t chase data perfection. We chase directional truth.
Why It Still Works (If You Build It Right)
Here’s the secret: most of our best models don’t use the “cleanest” data. They use the most honest data — even if it’s messy.
Because what you want, as an investor, planner, or operator, isn’t false precision. You want to know:
Is this project resilient?
Are we missing critical risks?
Where is the leverage to drive real change?
For those questions, context > completeness.
The best sustainability model isn’t the one with the fewest gaps.
It’s the one that keeps learning.
That’s what we aim for: adaptable systems, not fragile ones.
The Bottom Line: Humble Models, Better Decisions
We’re not modeling the universe. We’re modeling human systems — in flux, under pressure, imperfectly measured.
Expecting perfect data is like expecting a hurricane forecast to be exact to the street. What you want is clarity, movement, and enough signal to make a decision.
In sustainability, the numbers will never be perfect.
But if we keep our feet on the ground of reality, the insights will still be valuable.
Wangari’s Curated Reads
Sarah Gulley has some sustainable fashion resources you'll actually use. Her beautifully curated guide offers a thoughtful entry point into sustainable fashion with honest recommendations, from hard-hitting documentaries to hopeful podcasts and insightful books. Grounded in the spirit of Fashion Revolution Week, it reminds us that ethical fashion isn’t just about fabrics or trends—it’s about people, power, and accountability.
Fairytales From Ecotopia has some Perfect Consideration for you. The essay explores the intersection of Native American spirituality, legal resistance, and environmental sustainability through the lens of peyote use and the Native American Church. For our readers, it offers powerful insight into how indigenous philosophies and entheogenic traditions not only challenge extractive systems but also present an alternative, regenerative worldview grounded in sacred reciprocity with nature. It’s a call to rethink sustainability—not as efficiency or policy, but as a radical reimagining of our relationship with the Earth.
Metrics and Measurement remain the language of sustainability, writes Sowmy VJ. To understand how sustainability becomes action rather than aspiration, one needs to break down the often murky world of KPIs and reporting standards, and show how clear metrics can turn climate intentions into measurable progress. As African companies navigate ESG pressures and green finance opportunities, this piece underscores the importance of choosing meaningful data—and using it to drive both transparency and impact.
Thanks for the mention. Yes, metrics and measurement is where sustainability lives, and the right metrics drive the right kind of action.