That's not offending here because everyone is either working on a social or a business problem. All this as a part of our curriculum, we study design thinking, problem solving etc. This is a precis I had written on one of the problems I am working on. Posting it here for the good of the world!
Problem: Balancing the Bike Sharing System
We have bike sharing systems everywhere. These systems mainy allow people to rent/share/use bicycles in a defined space. These could be on campusses, work areas, SEZs or even certain areas of a city or a town. Usually there are stations from where the user can pick up a bike to travel to another station where he can drop it. Some of these bike sharing systems are free and some are paid. The major challenge in maintaining such bike sharing systems is to balance them or to make sure every user gets a bike when they need one. However, this challenge can be overcome upto some extent in bike sharing systems that require a payment or authorization from the user. Here, there is some data available on the no of bikes at different stations, who takes it and where they drop them. This data can help create a lot of patterns on bike usage based on weather, locality, person, events etc. Though this doesn’t solve the problem, there’s at least more data that can be used to create a solution.
But in case of Bike sharing systems which are free to use, there’s no data available and extracting this data using sensors can be expensive. In most of these systems basic data like the no of people using this system on a regular basis is also not available. This makes the problem more complicated. This exact problem is seen in the bike sharing system we have on our campus.
The bike sharing system on our campus is not completely balanced. At a point of time. nearly 100 bikes are in use through 5 bike stations. Out of these 5 stations, some always have insuffiecient or no bikes at all and some have too many at certain points of time in a day. The stations being linear makes the solution of finding a bike at a near by station also not feasible. If we can get more data on availability of bikes at stations, patterns can be analyzed based on weather, terrain, schedule etc giving more options to optimize the system may be by restricting the routes, or by setting up more stations or even physically transporting bikes from surplus stations to defecit stations.
What’s the most economical way to solve this problem? Can this problem be solved without collecting bike data?