Below is the plan of action I have laid down to have project through to completion.
Week 1 - Have a granular look at the data set to get the high level understanding on how to proceed with related work which i am going to deliver before the deadline.
Mobility of bike inside the city can be tracked by the sensor network like functioning of bike sharing system and this gives proper data for predicting the results in good manner. This project will focus on predicting bike demand based on different variables present in the dataset as well as different predictions will be evaluated based on the historical patterns for better management of bike sharing system by getting the user counts in order to keep the bike count in proportion with user counts.
I have included a synopsis of data set in issues section and how i aim to evaluate data for predictive analysis.
Week 2 - Hypothesis Testing
Having understood the data set, i believe crucial step for the success of bike sharing system is getting the information about fluctuating bike usage demand and arrange them in those stations where more users are coming and moving them from usage locations.
The Hypothesis Testing is now complete with my findings evaluating different factors influencing the bike rental along with R.code uploaded on to the repository.
Week 3Implementation
Getting good patterns from historic data is challenging task in moving vehicles at different locations and this can be achieved with different statistical as well as some data mining models.
I researched related work on the bike sharing system where scholarly individuals had utilized statistical model approach such as Shaheen, S., Guzman, S. and Zhang, H., 2010 proposed statistical based models for predicting the positioning of bikes at appropriate locations to avoid user dissatisfaction. I will be discussing more in the Related work section of my final report as it is influential work and helped me put ideas together for data analysis.
I aim to present a modeling approach to the problem, which has different data mining models such as Random Forest ,Neural networks ,PCA( Principal Component Analysis) and Conditional trees to predict the user counts which have registered and casual users who travel by bike sharing bicycles.
Future Weeks- I aim to set milestones for work i wish to accomplish which will be mainly implementation of different data mining models and algorithms for Kaggle submission.
I would seek help from supervisor as the issues arise and post up issues to for help on ongoing basis.
Below is the plan of action I have laid down to have project through to completion.
Week 1 - Have a granular look at the data set to get the high level understanding on how to proceed with related work which i am going to deliver before the deadline.
Mobility of bike inside the city can be tracked by the sensor network like functioning of bike sharing system and this gives proper data for predicting the results in good manner. This project will focus on predicting bike demand based on different variables present in the dataset as well as different predictions will be evaluated based on the historical patterns for better management of bike sharing system by getting the user counts in order to keep the bike count in proportion with user counts.
I have included a synopsis of data set in issues section and how i aim to evaluate data for predictive analysis.
Week 2 - Hypothesis Testing Having understood the data set, i believe crucial step for the success of bike sharing system is getting the information about fluctuating bike usage demand and arrange them in those stations where more users are coming and moving them from usage locations. The Hypothesis Testing is now complete with my findings evaluating different factors influencing the bike rental along with R.code uploaded on to the repository.
Week 3 Implementation Getting good patterns from historic data is challenging task in moving vehicles at different locations and this can be achieved with different statistical as well as some data mining models.
I researched related work on the bike sharing system where scholarly individuals had utilized statistical model approach such as Shaheen, S., Guzman, S. and Zhang, H., 2010 proposed statistical based models for predicting the positioning of bikes at appropriate locations to avoid user dissatisfaction. I will be discussing more in the Related work section of my final report as it is influential work and helped me put ideas together for data analysis.
I aim to present a modeling approach to the problem, which has different data mining models such as Random Forest ,Neural networks ,PCA( Principal Component Analysis) and Conditional trees to predict the user counts which have registered and casual users who travel by bike sharing bicycles.
Future Weeks- I aim to set milestones for work i wish to accomplish which will be mainly implementation of different data mining models and algorithms for Kaggle submission.
I would seek help from supervisor as the issues arise and post up issues to for help on ongoing basis.