1. Download the capitalbikeshare tripdata for the year of 2017 at the link
https://s3.amazonaws.com/capitalbikeshare-data/index.html which consists of 4 csv files.
2. Every student should submit a Jupyter Notebook for questions 1 and 2 by October 17, 2020 midnight.
3. The questions 3 and 4 should be worked in groups. Groups with 3 persons will be created randomly and
announced on October 9, 2021 in class. Each group should submit a single Notebook for questions 3 and 4
by November 6, 2021 midnight.
4. Penalty for late submission is 10%.
Questions:
1. Rental Numbers (Individual Work)
a. Calculate for every day hourly rental numbers by member types.
b. Explore the hourly rental patterns for all member types on different days.
c. Summarize your observations.
2. Rental Duration (Individual Work)
a. Write down your expectation on what type of member rents bikes for a longer duration.
b. Explore the relation between the rental duration and member types.
c. Interpret your findings.
3. Effect of the weather (Group Work)
a. Scrap the weather information for Washington DC from the website
https://www.wunderground.com/history/daily/us/dc/washington/KDCA for the year 2017. You may find
selenium package very useful for this task. You can delete the ‘Wind’ feature from the dataset. It won’t
be needed. It will take around one hour.
b. Return the dataset into an hourly time series.
c. Find the most important 3 weather features that effects the total (casual+member) number of daily
rentals.
d. Explore the hourly rental numbers on different weather conditions for all member types.
e. Find an abnormal day (that is rental numbers do not follow the rental patterns you have explored in
question 2b). Explain how you found that day.
4. Busiest Station (Group Work)
a. Give your definition of the “busiest station”.
b. Based on your definition, find the busiest stations for every month and for every member type.
c. Do the stations you found make sense? Explain. For this question you need to have some information
on the city Washington DC.
