The tourist information offices in both Arosa and Lenzerheide staff their offices according to how many tourists they expect to be contacted by. The process of staffing the office relies on experience and laborious mapping of the next months events, season, holidays etc.

Customer requirements: The tourist office would like a tool to facilitate and enhance the prediction of tourist requests.

Potential Key Deliverables: Forecasting tool one month prior.

Out of Scope: Predictions for the sub-branch offices/ information points.

Potential Challenge: A prediction tool should forecast the amount of tourist requests throughout different channels such as phone calls, emails, and front desk per month. The prediction depends on a variety of factors including events, school holidays, public holidays, season and weather. The importance of these factors should be explained.

Results

Details on the implementation of the project during the hackdays can be found here. The Lenzerheide files are in the repo "forecast5.0" and the Arosa ones in repo "forecast5.1".

You can find the final presentation as well as a demo of our forecast here.

Enjoy!