#forecasting is an active hashtag on Bluesky. In the last 30 days, 25 people shared 72 posts with it — around 2 a day. Activity is up 67% versus the previous week, peaking on Aug 31 with 10 posts.
Tags most often used together with #forecasting.
www.theatlantic.com
Hurricane Science Was Great While It Lasted
The U.S. is hacking away at support for state-of-the-art forecasting.
otexts.com
Forecasting: Principles and Practice, the Pythonic Way
arxiv.org
Responsible forecasting: identifying and typifying forecasting harms
Data-driven organizations around the world routinely use forecasting methods to improve their planning and decision-making capabilities. Although much research exists on the harms resulting from tradi...
apnews.com
The government cuts key data used in hurricane forecasting, and experts sound an alarm
Weather experts are warning that hurricane forecasts will be severely hampered by the upcoming cutoff of key data from U.S. Department of Defense satellites.
openforecast.org
The Menace of ML: Simple Moving Average - OpenForecast
Is the Simple Moving Average a good forecasting method? What evidence says, how to choose the order, and when to move to better models.
gumroad.com
Forecasting Metrics That Don’t Lie - Core Edition
📘 FORECASTING METRICS THAT DON'T LIEChoosing, Testing, and Monitoring Forecast Metrics for Demand, Inventory, and Risk DecisionsValery Manokhin, PhD, MBA, CQF · 310 pages · 12 chapters · Python throughoutA model that wins on MAPE can lose money on the shelf. A 15% accuracy gain can hide a 12% bias. An interval that looks tight can be wrong a third of the time.Forecast metrics are decision instruments, not neutral truth detectors — and most teams are still choosing them by habit. This is the complete reference for forecast evaluation: what each metric actually elicits, what it rewards, where it breaks, and what to use instead.Why this book-------------Most forecasting books spend one chapter on evaluation. This one spends twelve. Every metric arrives with its assumptions stated, every criticism carries a remedy, and every chapter opens with a failure that really happens in production — the forecast that won on paper and lost in production, the spare parts disaster, the reconciliation illusion, the model that rotted silently.A retrospective Walmart M5 case study runs through the whole book, so the same data is re-examined as the metrics get more demanding. Its cohort-selection limitations are stated explicitly rather than glossed over.What's inside — all twelve chapters, complete---------------------------------------------------Part I — Foundations of Honest Evaluation✓ Why metrics mislead, the three types of drift, and a decision-centred taxonomy that replaces the usual alphabet soup✓ Temporal validation, naive baselines, and statistical tests for whether a difference is real at all✓ MAE, MSE and RMSE through elicitation — the median versus the mean, objective mismatch, and ranking reversals across metricsPart II — Scale, Bias, and Intermittent Demand✓ MAPE's fatal asymmetry, why sMAPE made it worse, and where WAPE and MAAPE actually help✓ MASE, RMSSE and WRMSSE — including the denominator trap that quietly invalidates published comparisons✓ RMSSE-B: an author proposal for scaling against the business benchmark rather than a statistical one✓ Bias metrics, tracking signals, the bias–accuracy decomposition, and Forecast Value Added for judging whether human overrides earn their keep✓ Intermittent demand in full: ADI and Syntetos–Boylan classification, Periods in Stock, lead-time demand, SPEC, and Croston/SBA/TSBPart III — Distributions, Dependence, and Decisions✓ Calibration and sharpness, PIT histograms, reliability diagrams, pinball loss, CRPS, interval and Winkler scores, log score✓ Formal calibration tests (Kupiec, Christoffersen) and conformal prediction's coverage guarantee under exchangeability✓ Energy and variogram scores, coherence metrics, temporal hierarchies, copula-based evaluation✓ The accuracy–utility gap: ranked probability score, information ratio, newsvendor cost and regret, service levels and fill ratePart IV — Production Evaluation Systems✓ Drift and failure modes, monitoring with PSI and CUSUM, forecast stability and revision volatility✓ Alerting thresholds, triage from alarm to diagnosis, retraining triggers✓ Metric bundles, scorecards, domain bundles for supply chain, energy and finance, evaluation governance, and a maturity modelWhat you get------------✓ The book — 310-page PDF, 7×10, fully indexed and cross-referenced✓ Companion source — Python code, data-acquisition instructions, and reproducibility checks for selected numerical and semantic claims✓ Overleaf source — the full LaTeX project✓ A metric specification sheet — the book's single sign and denominator contract, so a disagreement about a sign has one place to be settledYou obtain the M5 source data separately under its own terms.Who this is for---------------✓ Data scientists and ML engineers who need evaluation that survives review✓ Demand planners and analytics leads in retail, manufacturing, finance and energy who make operational decisions from forecasts✓ Researchers, analysts and students who want one authoritative reference on forecast accuracyFormulas are given in full, but the book is written for people who have to justify a metric to a business, not only to a journal. No heavy maths background required.------------------------------------------All twelve chapters are written and shipping today. This is an actively maintained book — corrections, new case studies and new material arrive automatically at no extra cost, for life.------------------------------------------------------------🔥 Measure what matters. Stop being fooled by metrics.https://valeman.gumroad.com/l/forecasting_metrics
freegardner.com
Almanac Weather Claims vs Scientific Forecasting
Almanac Weather Claims vs Scientific Forecasting
spaisee.com
Google’s WeatherNext 3 Tests the Business Case for Faster AI Forecasts
Google’s WeatherNext 3 adds raw satellite observations to speed AI weather forecasts, but its commercial promise depends on improving local accuracy during storms, heat waves and other rapidly changing conditions.
Posts are pulled live from Bluesky and cached briefly. Posts with content labels are hidden.