PredictHQ Demand Impact Pattern Makes Severe Weather Events Consumable and Explainable for Demand Forecasting

By AIT News Desk On Nov 8, 2021

PredictHQ announced its Demand Impact Pattern for severe weather events, with data sets to help businesses prepare for major weather events and mitigate overall impact by integrating into machine learning models for demand forecasting. Research shows that abnormal weather disrupts the operating and financial performance of 70% of businesses worldwide. Every year, weather variability is estimated to cost $630 billion for the U.S. alone. Built specifically for the retail industry, PredictHQ’s Demand Impact Pattern gives data scientists a way to improve their forecast accuracy by incorporating weather event data into machine learning models, providing customers visibility into the leading, day-of, and lagging effects of severe weather events.

PredictHQ’s Demand Impact Pattern helps customers better understand exactly where and when a severe weather event is taking place, including the expected impact repercussions to the lead up of an event, the actual event date, and the aftermath in the days after. This enables faster action and saves retailers millions in revenue and operational costs.

New research from PredictHQ analyzed thousands of retail locations of large, national quick serve restaurants (QSRs) to identify the generalized impact on retail for 73 severe weather event types. PredictHQ found 58% of stores had significant forecast accuracy improvement incorporating severe weather due to:

Identifying and accounting for previous historical anomalies to be excluded from baseline modelling
Increasing a forecasting team’s understanding of when a severe weather impact would begin, and how long it would last for

The company also today announced the availability of Polygons, which provide accurate details of a geographic area impacted by a weather event. Both Polygons and Demand Impact Pattern enable teams to have data-driven strategies matched to a severe weather event’s intensity.

Pacing Innovation in AI Development: Avoiding the Pitfalls of Early Launches

Nov 15, 2024

Enhancing Telecom Networks for Efficient AI Workload Management

Nov 15, 2024

Tessl Raises $125 Million to build AI Native Software Development

Nov 15, 2024

Prev Next 1 of 40,944

PredictHQ’s Demand Impact Pattern helps data scientists and executives:

See the full picture of a severe weather event so they can properly understand total impact before, during, and after the event
Conduct tailored, store-level forecasts that help to decide when or if to pull a location offline and service an area through other locations, and reallocate resources in real-time
Adjust staffing needs, update inventory plans and orders, redistribute assets to stores that will be less affected, and prepare targeted recovery strategies

PredictHQ’s Polygons Feature:

Provides granular, geographic information, rendered as model-ingestible data so forecasting models can evolve as quickly as severe weather watches and warnings are announced
Enables data teams to more accurately search and pinpoint actual geographic areas affected by a severe weather event via an easy to view map interface and intuitive map search function
Helps retailers understand which stores will be immediately impacted by a weather event and which supply chains will be disrupted

“You can’t control severe weather but you can control your staffing and stocking to minimize impact and save millions each year,” said Campbell Brown, CEO at PredictHQ. “Previously, data scientists haven’t had the ability to make rapid, accurate updates to demand forecasts and plans, because there has been no intelligence layer making sense of severe weather events — we’ve changed that today. In the age of climate change, all businesses are seeking to get better at managing this impact as it rapidly escalates, and this is the first set of tools to make severe weather data fit-for-purpose in retail demand forecasting updates.”