Two Types of Weather Indexes for Business Performance

In order to deduce the relationship between weather and business performance, I am researching different literatures to find the solutions being used. Here are two types of solutions to present the relationship: (1) Identify the current weather pattern to explain the relationship to the business and (2) develop a regression model to include dependent variables to calculate the impact to business.

Synoptic weather pattern analysis with business performance:

Brown et al. in 2019 investigated different weather variables (including temperature, precipitation, and weather pattern) to calculate in regression and collaborate coefficients to find the relationship between weather factors and business performance of bike hiring. They used the synoptic charts of 30 groups developed by Met Office in 150-year records to classify the current weather. Based on the characteristics of each weather group, they can explain the weather impact to the business. For example, during the low temperature and strong wind, the total number of journeys reduced a lots from the normal weather.

30 weather patterns (Fig. 2 in Brown et al. (2019))

The time series of total journey and weather pattern number to show their relationship (Fig. 8 in Brown et al. 2019)

Regression model with weather and business factors:

Roth Tran in 2018 develop a weather index via LASSO regression model in residuals-to-residuals framework. By the weather index, we can use weather variables (like temperature, precipitation etc.) and business variables (like the monthly marketing event, holiday events etc) to predict the change of net sales. The author noted that based on different products the same index may represent different weather status. For example, soup sale may increase when temperature drop, but ice cream sale increases when the temperature go up.

The baseline equation for the weather index (Eq. 1 in Roth Tran (2020))

Weather index changes by the temperature in different store environments (Fig. 8 in Roth Tran (2018))

These two types solutions came from different research domains: Synoptic weather pattern analysis is a classical meteorological analysis to explain the weather situation; the regression model is more related to economical analysis which can quantify the contributions from different variables. When we develop an application from weather factors on the economics, we may need to consider how to collaborate both solutions.

References:

Brown, H., Lee, M., Steele, E., Neal, R., & Chowienczyk, K. (2019). Using weather sensitivity analysis to predict business performance. Weather, 74(7), 231-236.

Roth Tran, B. (2018). Taming Thor: A new approach to modeling weather effects. Available at SSRN 3464678.

Roth Tran, B. (2020). Sellin’in the Rain: Adaptation to Weather and Climate in the Retail Sector. Available at SSRN 3337110.