Using historical data
from the past 3 to 5 years, we calculate if and in what way the weather can affect a
company´s sales and demand for its products.
Regression analysis
Using a standard statistical method known as regression
analysis, costumer-specific ratings are compared with data from SMHI´s database. The
input data can be structured on a daily basis, per week or month, and it may represent
sales or demand in a given area, in a region or for a whole nation.
Total weather picture
The result of the regression analysis indicates just how strong an effect the weather has
on performance. This rating is given with correlation coefficients, where +1,0 or -1,0
indicates 100% co-variation and where 0 shows that the weather has no effect whatsoever.
The effect of the weather is tested with regard to various individual weather parameters,
but it often shows the strongest links when we run a combination of several weather
parameters simultaneously in a process known as multiple regression. The sensitivity
analysis also includes clearing the costumer-specific statistics from all factors not
affected by the weather.
Weather Sensitivity Index Forecast
In many industries, we have been able to use the
sensitivity analysis to verify strong weather-related influences. The correlation
coefficient is often between 0,65 and 0,85, sometimes even above 0,90. With such strong
links, we suggest the company subscribe to the Weather Sensitivity Index Forecast on a
regular basis, whereby the company receives reliable information via the regression link
as regards such factors as expected sales in the forthcoming 10-day period.
SMHI Weatherwise undertakes weather sensitivity analysis
for countries in central and northern Europe. |
