In real world, at citi bike pick up stations sometimes it happens that if there are 50 bikes at a particular station and if the pick up time is the peak time of office and people who want citi bike is more than 50 then there is loss of customers. And sometimes if it is not the peak time and there are 30 bikes but for pick up we have only 10 people then there is again loss. So by analyzing the dataset using various factors like temperature, humidity, condition, windspeed, visbility, datetime we can predict the pick up demand and on the basis of predicted value we can keep the required number of bikes at particular station at a specific time.