Lakehouse architecture
Demand, capacity and cash forecasting on DataNature: build forecasts that teams trust
DataNature supports forecasting from data to decisions. Historical sales, usage, load or cash flows land in Iceberg tables; dbt builds clean time series with calendars, prices and events; Spark trains many forecasts in parallel with experiment tracking; the model registry versions them; scheduled jobs write forecasts and prediction intervals to Gold tables; and dashboards, AI360 Studio and Voir put them in front of planners. Forecast accuracy is tracked against actuals so models improve over time.
Published . Last updated
On this page
Every organization forecasts something: product demand, network capacity, energy load, call volumes or cash. Forecasts are only useful when planners trust them, which means consistent data, honest uncertainty and visible accuracy. This guide shows how to build that on DataNature.
What can you forecast?
| Industry | Forecast | Used for |
|---|---|---|
| Retail and CPG | Demand by product and store | Replenishment and promotions |
| Telecom | Traffic by cell and region | Capacity planning and investment |
| Energy and utilities | Load and renewable generation | Grid balancing and trading |
| Banking | Cash demand, deposits, liquidity | ATM cash, treasury and ALM |
| Contact centres | Call and chat volumes | Staffing |
| Manufacturing | Demand and material needs | Production planning |
How do you build forecasts, step by step?
- Assemble the time series: actuals at the right grain (daily by store and product, hourly by cell) in Gold tables, with gaps and outliers handled in dbt.
- Add drivers: calendars, holidays, prices, promotions, weather, events and outages as features.
- Choose methods by series: simple statistical models for stable series, gradient-boosted or global machine learning models across many related series, and hierarchical reconciliation so store forecasts add up to region totals.
- Train at scale: Spark trains thousands of series in parallel; experiment tracking records methods, parameters and backtest accuracy.
- Backtest honestly: rolling-origin evaluation over past periods, with error measures such as MAPE or weighted error that planners understand.
- Publish forecasts with intervals: scheduled jobs write point forecasts and prediction intervals, with model version, to Gold tables.
- Close the loop: compare forecasts with actuals every period, show accuracy by segment, and retrain when accuracy drops.
How do planners use the forecasts?
- Dashboards in BI and certified cubes in the Semantic Layer show forecast, actuals and accuracy in one place, in Power BI or Superset.
- Planners can adjust forecasts with recorded overrides, and the platform measures whether overrides helped.
- Voir agents answer questions such as "why is next week's forecast for store 12 higher?" with the drivers behind it.
- Forecasts feed replenishment, staffing or capacity decisions through reverse ETL.
Which use cases are ready to start from?
- Demand forecasting and replenishment in retail
- Capacity planning in telecom
- Load and renewable generation forecasting
- Liquidity and ALM analytics in banking
- Demand forecasting and S&OP in manufacturing
Frequently asked questions
Do we need deep learning for forecasting?
Rarely at first. Well-built statistical and gradient-boosted models with good drivers are strong baselines; add deep learning only if backtests show a clear gain.
How often should forecasts refresh?
As often as decisions are made: daily for replenishment, hourly for grid or network operations, monthly for budgets.
Sources
How does DataNature Atlas keep this secure and compliant?
Governance is part of the architecture, not a separate tool bolted on later. DataNature Atlas is the platform's governance module:
- Knows your data: a catalog with automated PII detection and classification across Bronze, Silver and Gold.
- Controls access: role-based grants plus attribute-based policies, row filters and column masks (null, redact, partial, hash, year only).
- Governs AI: Voir agents are scoped principals; masks, row filters and PII guardrails apply to every tool call and answer.
- Proves it: lineage from source to metric to decision, and an audit log of who or what accessed which data.
- Stays local: Atlas runs inside your deployment, so policies, lineage and logs never leave your jurisdiction.
These controls support GDPR, PIPEDA, Quebec Law 25, CCPA, HIPAA and sector rules such as OSFI E-23. Read DataNature Atlas in depth or see security and compliance.
Platform components in this article
Learn more about each DataNature product and module mentioned here.
Products
See it on your own data.
Book a demoDemos and quotes are handled by Dlytica, the company behind DataNature.