DESCRIPTION
The purpose of the experiment is the creation of a Reccomendation Enegine for e-bike rental which could do the following actions:
1) Forecast the number of reservation in a specific rental point depending on the weather forecast and so recommend other destinations in case of bad-weather or overbooking
2) Reccomend to the user interesting things to do based on the existing partners and points visited so far. Reccomend interchange with the train at local train stations in order to avoid car trips.
I-Space Coach
![]() |
I-Spaces Involved
![]() |
Create a Reccomendation Enegine for e-bike rental which could do the following actions:
1) Forecast the number of reservation in a specific rental point depending on the weather forecast and so recommend other destinations in case of bad-weather or overbooking
2) Reccomend to the user interesting things to do based on the existing partners and points visited so far. Reccomend interchange with the train at local train stations in order to avoid car trips.
Using weather information and positions of the e-bikes in order to forecast and suggest solutions to the users, in order to reduce their impact on the environment
We aim at using data to promote a zero-pollution choice for the tourists and increase the attractiveness of e-bikes+train instead of
cars.
The basic idea is to provide a solution to travel to rural areas with zero-pollution. The user who wants to visit a destination can get
the recommendation to get to a train station where he or she can pickup an e-bike and reach his/her final destination with no
impact.
The recommendation engine will use the historical data about reservations, anonymized routes of the users and weather forecasts to increase the probability to make in interesting proposal and so generate a green travel.
Define an algorithm which combines the weather information with the historical data about reservation and allow a reliable prediction in the e-bike market.
Create a reccomendation engine which will propose what to visit in a specific destinations
1) forcast number of correctly predicted cancellations (85% accuracy at the end of the experiment)
2) 25M of GPS data processed by the reccomendation engine