Using weather sensitivity analysis to predict business performance
Author(s) -
Brown Hannah,
Lee Malcolm,
Steele Edward,
Neal Robert,
Chowienczyk Katie
Publication year - 2019
Publication title -
weather
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.467
H-Index - 40
eISSN - 1477-8696
pISSN - 0043-1656
DOI - 10.1002/wea.3581
Subject(s) - sensitivity (control systems) , regression analysis , computer science , set (abstract data type) , statistical analysis , meteorology , data set , weather forecasting , data analysis , numerical weather prediction , environmental science , statistics , data mining , engineering , geography , mathematics , machine learning , artificial intelligence , electronic engineering , programming language
For many businesses, the weather is a strong driver of performance. Here, we introduce two assessment tools for organisations wanting to identify these links, so that they may subsequently be used to predict future changes. By combining data collected by businesses with historical weather data, the assessment tools use a set of pre‐defined statistical analysis methods to quantify their particular sensitivities. This analysis is fast and flexible, to facilitate the ease of incorporation of meteorological effects into the corporate decision‐making process. The weather sensitivity analysis conducted includes both correlation and regression analysis and weather pattern analysis, providing results suitable for subsequent operational application within a real‐time forecast system. A demonstration is provided for bike hire data collected under the Santander Cycles scheme, published by Transport for London, wherein it is shown that 74% of the variability in the journeys undertaken may be explained by a simple statistical model involving only two weather variables (temperature and rainfall).