PROJECT FINAL PAPER

EE3: Weather, Weekdays, and Bike Rentals: A Quantitative Analysis

Huiling Liu, Jiaying Zhou, Min Jiang, Shreya Sawant, Yi Kiu Ho, Zuo Wang

CIAM

BUS501

Dr. Edmund Khashadourian

July 12, 2025

Based on EE3 definitions, we did begin with data cleaning and organizing our data before it in the same manner by converting temperatures from Fahrenheit to Celsius for better handling. Data cleaning is important because it can lead to greater reliability and validity in making decisions and analyzing data. Hermanson et al. (2021) explain that validity, completeness, and accuracy in supporting data set the quality of insights. Furthermore, information is a strategic asset that has the potential to offer a valuable competitive advantage (Vasarhelyi et al., 2023).

Solving the Missing Data Problem

The data set did not have the maximum, minimum, and average temperature on February 1 to February 14, 2023. To complete the data set, one had to find out the geographic location where the data collection would be in a position to obtain the right weather details. The data set was not specific about the location. We employed AI-driven analysis to identify the location, which was Washington, D.C. The missing temperature figures in the data set were eventually obtained from places of pleasant climate to fill the gap.

Weather and Bike Usage Analysis

One can easily observe in the graph that the use of bikes drops significantly during rainy seasons. The evidence proves that people do not like to drive when it rains. This comes in handy for preliminary analysis when a company is deciding whether it should venture into a new market whose weather might affect total sales. Research shows that extreme weather events—such as snow and heavy rain—considerably hinder transportation infrastructures (Skevas et al., 2025). Since cycling is highly weather-dependent, weather conditions like rain and icy roads can drastically limit cycling use.

Additionally, weather condition data reveals that "partly cloudy" days are the most common type of weather in the two-year period under observation, and also the days on which maximum usage of bikes is noted. This again supports the hypothesis that weather has a considerable bearing on usage.

Casual vs. Member Bike Usage

It is clear that members utilize bicycles significantly more than non-members—nearly twice as much. This shows that members use the bike-sharing system more since they use bicycles for their daily transportation or as their day-to-day transport.

Weekday vs. Weekend Rentals

The weekday rentals totaled approximately 11,900,000 and weekend rentals totaled approximately 4,860,000. The weekday daily average rentals totaled approximately 21,046 and weekend daily average rentals were 23,307.

To ascertain whether this variation is significant, we used a t-test. Likewise, since the determination of e-cargo bike freight demand is made using quantitative models, we used a t-test to ascertain whether variation between bike use on weekend and weekday is significant. This enables fact-based decision-making and not assumption-based (Mantecchini, Nanni Costa, & Rizzello, 2025). The t-test result gave approximately a t-statistic of 3.32 and a p-value of ~0.0010, which shows there existed a statistically significant difference in bike usage on weekdays and weekends. The result shows people are likely to use bicycles more during weekends than weekdays.

Bike Type Preference

Statistics indicate that classic bikes are the most utilized. That implies that users will choose classic bikes more often, which is most likely due to the fact that most people don’t try or adopt the new trend of the eclectic bikes. Because the data geographically are centered in Washington, D.C., which is a relatively small city compared to other cities like Los Angeles, it would logically follow that normal bikes are utilized more with the relatively shorter distance.

References

Hermanson, D. R., Lawson, J. G., & Street, D. A. (2022). Detecting and Resolving 'Dirty' Data: Certified Public Accountant.  The CPA Journal, 92(7), 36-41. https://2q21eenab-mp02-y-https-www-proquest-com.proxy.lirn.net/scholarly-journals/detecting-resolving-dirty-data/docview/2708409756/se-2

Mantecchini, L., Nanni Costa, F. P., & Rizzello, V. (2025). Last Mile Urban Freight Distribution: A Modelling Framework to Estimate E-Cargo Bike Freight Attraction Demand Share.  Future Transportation, 5(1), 31. https://2q21eeas6-mp01-y-https-doi-org.proxy.lirn.net/10.3390/futuretransp5010031

Skevas, T., Thompson, W., Brown, B., Salin, D., Gastelle, J., & Edgar Marcillo-Yepez. (2025). Weather extremes and their impact on crop transportation networks: Evidence from U.S. Midwestern elevators.  PLoS One, 20(3) https://2q21eendo-mp02-y-https-doi-org.proxy.lirn.net/10.1371/journal.pone.0319815

Bike Usage Weekend vs. Weekday

Total Sunday Monday Tuesday Wednesday Thursday Friday Saturday 2272348 2194266 2384470 2485660 2484340 2436002 2576308

Sum of Total Bike Usage precipitation versus no precipitation

Total No Yes 10993934 5641866

Precipitation

Weather condition and Bike Usage

Total Clear Overcast Partially cloudy Rain Snow Snow, Partially cloudy 1158332 340876 9494726 5569526 64744 7596

Caual vs. Member

member casual 5204474 3212223