Data Mining for Business
Data Mining for BusinessTeam C Project Executive SummaryBYGB 7967 004December 2, 2018 Hanwen FengQiaoye ZhangJohn DeLucaZixin Wei        AbstractThe bike-sharing rental process is highly correlated to the environmental and seasonal settings. By processing the dataset from Capital Bikeshare System in Washington DC, we use neural networks, decision trees, and clustering to figure out if there is a correlation between pertinent transportation factors and the number of uses. From our model, we see that factors such as temperature, relative humidity, precipitation and month are most likely to dictate the usage of bikes. From the company’s perspective, we should be able to determine the number of bikes to be used based on weather forecasts, which should capitalize on days of high demand and decrease depreciation on low demand days, making them last longer.IntroductionOur topic is capital bikeshare system in Washington D.C. Automatic bike-sharing systems have expanded their appearance all over the world in the past decade. These services allow users to have quick access to rent bikes for a short period of time and return them at any rental stations with a small fee charged. Its existence not only adds an economic alternative to public transportation but also help to further reduce city congestion.
Currently, there are about over 4,300 bikes and 500+ stations in D.C.. The duration of travel, departure and arrival position is explicitly recorded in the system. This feature turns bike sharing system into a virtual sensor network that can be used for sensing mobility in the city. Hence, it is expected that most of important events in the city could be detected via monitoring this data.Data DescriptionOur original dataset is from UCI Machine Learning Repository with 17379 records. And we got hourly temperature data from frontierweather.com with 17545 records. Merged these two, we got our final dataset to do the analysis.  Problem statementThe bike-sharing system is revolutionary to traditional bike rentals where the entire process from membership, renting, and returning has become automatic. Through these systems, users are able to easily rent a bike from a particular location and return to a different station. Currently, there are about over 500 bike-sharing programs around the world which are composed of over 500 thousand bicycles.The target group for our research is the Capital Bikeshare company in Washington D.C. We are analyzing their total rider dataset from a two-year period examine the optimal number of bikes that should be made available around the city given the pertinent weather measures on a given day. Various external factors such as temperature, precipitation, the day of the week, season, the hour of the day, etc., can affect users’ rental behaviors. Our model attempts to capitalize on days of high demand to increase profit while reducing depreciation costs for the bikes and overall bike system during low usage days.