Skip to content
Noether Analytics

Mathematical models for operational business decisions in your environment.

Your data. Your constraints. Your systems. We build machine learning and optimization models around the way your business works.

FROM COMPLEXITY TO A DECISION
01 / THE STARTING POINT

The problems we solve.

Some decisions are too complex for a spreadsheet. Others take more manual work than they should. Does any of this sound familiar?

Your problem isn't on the list? Tell us what you are trying to decide.

Already using planning software? We start where its rules, forecasts or workflows reach their limits.

Manufacturing · Wholesale & retail · Logistics · Construction · Energy · Financial services

02 / OUR METHODS

Two methods. One working solution.

Predicting what might happen and deciding what to do are different problems. We bring the right methods to each, and connect them where it helps.

ŷ = f(x) with an 80% prediction interval
01 / ESTIMATE

Machine learning

Understand what is likely to happen.

We learn patterns from your data to forecast demand, estimate risks, detect anomalies and interpret documents. We test on data the model has not seen and compare against the way you work today.

  • Forecasting
  • Risk & anomaly detection
  • Document understanding
max x₁ + 3x₂ subject to x ∈ F
02 / DECIDE

Mathematical optimization

Work out the best feasible action.

We model your objective, the decisions you can make and the constraints you must meet. Then we solve for a feasible plan, measure its quality and show the trade-offs.

  • Planning & scheduling
  • Resource allocation
  • Purchasing & inventory
FROM MODEL TO DAILY OPERATIONS

Built to work in your environment.

We connect to the data and tools your teams already use: ERP, databases, spreadsheets, documents and APIs. Deployment, access controls and human approvals are agreed around your requirements.

  1. 01Your data
  2. 02A model of your operation
  3. 03A decision your team can use
BACKTESTING

Measure the difference. Before you roll it out.

Before a rollout, we test forecasting and planning models on your historical data, using only information available at each decision point. We compare the results with your current approach using agreed metrics, so you can assess the potential for your operation.

  1. 01Your historical data
  2. 02Your current approach as a benchmark
  3. 03Results you can measure

Start with a specific decision. Test against current practice. Make it usable.

First we agree the scope, data and success criteria. A pilot compares the model with your current approach. Integration, ongoing operation and handover are agreed together.

03 / USE CASES

A model is useful when it changes a decision.

Explore seven ways this can look in practice. These are illustrative use cases with synthetic data.

DECISIONS IN FOCUS

Swipe to explore use cases

ILLUSTRATIVE USE CASE

Know what to plan for.

THE PROBLEM

Seasonality, promotions and sparse sales make demand hard to estimate. A single forecast number hides the risk of being wrong.

THE DECISION

A demand estimate and its uncertainty, ready for purchasing and capacity decisions.

Weekly demandInteractive example

Week +4, forecast

175 ± 13 units

80% interval: 162 to 188 units. 95% interval: 156 to 194 units.

ObservedForecast80% interval95% interval
Choose a week
Synthetic data. Model: trend and yearly season with a demand level that drifts over time, fitted on two years; the chart shows the last year. The further ahead the forecast, the further the level can drift, so the intervals widen with the horizon.
Explore this use case

Illustrative data. No client result or performance claim.

04 / ABOUT US

Meet the team.

We are three ETH Zurich graduates building models that are used in day-to-day operations. You work directly with the founders, from the first conversation to implementation.

05 / LET'S TALK

Which decision would you improve?

Tell us what takes too much time, costs too much or is hard to plan. We will work out together whether a model would help.

Or write to us directly

info@noether-analytics.com

A first conversation to understand the problem. No preparation needed.

We use your details only to answer your inquiry. See our privacy notice.

06 / CAREERS

We are hiring.

Enjoy mathematics. Build things that work. Join us in bringing machine learning and optimization into real operations.

Apply to Noether

Opens our application form in a new tab.