Drivers, regressors and signals are different words for the factors that help explain why museum visitor numbers go up or down. In VisitorCast, they are used to improve visitor forecasts by adding context to historical attendance data.
A museum does not receive the same number of visitors every day. Visitor numbers change because of weekends, school vacations, public holidays, exhibitions, events, weather, tourism and many other factors. These factors are the signals VisitorCast uses to understand visitor demand.
What is a driver?
A driver is a factor that may influence visitor numbers.
For example, a school vacation can increase visits from families. A major exhibition can attract more tourists and members. Heavy rain may reduce visits for some museums but increase visits for indoor attractions. A public holiday may create a peak if the museum is open, or no visitors if the museum is closed.
Common museum visitor drivers include:
- Weekends
- Public holidays
- School vacations
- Seasons
- Exhibitions
- Events
- Weather
- Ticket prices
- Local tourism
- Marketing campaigns
The word “driver” is useful because it is easy to understand for non-technical users.
What is a signal?
A signal is data that may contain useful information for the forecast.
For example, “Saturday” is a signal. “Rainy weather” is a signal. “School vacation in Denmark” is a signal. “A new exhibition opened this week” is also a signal.
Some signals are strong and reliable. Others are weaker. For many museums, day of week and seasonality are strong signals. Weather may be weaker, but still useful in specific situations. For example, bad outdoor weather may increase demand for some indoor museums but reduce demand for museums that depend on gardens, outdoor areas or long travel distances.
What is a regressor?
A regressor is the technical forecasting term for an external variable used by the prediction model.
In practical terms, a regressor is just a structured input that helps the model explain visitor numbers. If the model knows that a day was a public holiday, during school vacation, part of a major exhibition and a Saturday, it has more context than visitor numbers alone.
VisitorCast can use regressors to improve the forecast when those regressors are relevant and available for both the historical period and the future forecast period.
Weather signals in VisitorCast
VisitorCast offers two weather signals you can switch on for your museum from the museum settings page. Both are off by default.
Weather Score is a 0–100 measure of how pleasant a day is, combining temperature, sunshine and a penalty for rain and snow. Higher means more pleasant. For some museums pleasant weather lifts visits; for others it draws people outdoors instead — VisitorCast learns which way it goes from your own history.
Rain & Snow Score is a 0–100 measure of how wet a day is, where higher means more rain or snow. It is most useful for museums whose visits clearly rise or fall with wet weather.
Because Weather Score already includes a rain-and-snow penalty, the two signals overlap, so only one of them can feed your forecast at a time. To use either one, first pick a weather location (a nearby city) on the museum settings page — the signal label then shows that city, for example Weather Score, Copenhagen.
Weather is only used on the 7-day forecast. Reliable weather data is only available about two weeks ahead, so the 30-day and 90-day forecasts leave weather out.
Why do drivers matter?
Drivers matter because museums are affected by real-world behavior.
If visitor numbers increase every Easter, the forecast should know when Easter occurs. If an exhibition creates higher visitor numbers, the forecast should know when the exhibition starts and ends. If weekends are consistently busier than weekdays, the forecast should learn that pattern.
Without drivers, the forecast can still detect patterns in past data. But with drivers, the forecast can better explain why those patterns happened.
Strong and weak drivers
Not every driver has the same value.
Strong drivers are usually consistent and repeated. Examples include:
- Day of week
- School vacations
- Public holidays
- Seasonality
- Opening days
Medium drivers may matter, but depend on the museum. Examples include:
- Exhibitions
- Events
- Weather
- Ticket price changes
- Marketing campaigns
Weak drivers may be difficult to measure or only matter in specific cases. Examples include:
- Cruise ship arrivals
- Nearby city events
- Google Trends
- General tourism indicators
A weak driver is not useless. It simply means it should be tested and evaluated rather than assumed to improve every forecast.
How to choose the right drivers
Start with the drivers that are easiest to trust and maintain:
- Opening days and closed days
- Day of week
- Public holidays
- School vacations
- Exhibitions
- Events
- Weather
Then add more advanced drivers if they are relevant for your museum, such as tourism data, cruise ships, large nearby events or website visits.
The best driver is not always the most advanced data source. The best driver is the one that reliably explains visitor behavior and can be maintained over time.
Conclusion
Drivers, regressors and signals all describe the same basic idea: extra information that helps VisitorCast understand visitor demand. For most museum users, “driver” is the clearest term. For technical users, “regressor” explains how the data is used in the forecasting model. For product communication, VisitorCast can use both: “drivers, also known as forecasting signals or regressors.”