This section discusses findings about the relationship between bike share ridership and weather data for the city of Toronto that was retrieved from Meteostat.

<aside> 💡 All weather readings were obtained from the Toronto airport weather station. Bike share stations are situated at multiple locations across the city, which are not near the airport. There is expected to be some variation between conditions recorded at the weather station and those observed at the location of all the bike share stations across the network.

</aside>

🌡️ Daily Ridership and Maximum Daily Temperature

Average daily ridership is shown in the line chart below for all days between January 1, 2018 and December 31, 2022 in comparison to the range of daily temperatures which are shown in the grey shaded area

14_daily_max_temperature_and_ridership.png

Daily ridership peaks during the middle of the year, in agreement with monthly ridership seen earlier. Strong yearly seasonality is seen in both the daily temperature range (light grey shaded band) and the average daily ridership (dark and lighter green lines) at both types of stations.

The correlation between average daily ridership and maximum daily temperature is shown in the heat map below by year

15_daily_max_temperature_ridership_correlation.png

The heat map shows a strong correlation ($R^2$ > ~0.85) between daily temperature and daily ridership. This strong correlation is typical of bike share ridership across North America. Given the strong and matching yearly seasonality in daily temperature and daily ridership, it is not surprising that these correlations are so strong. For obvious reasons, the correlation dipped during 2020 due to the gap in ridership in March of that year. This gap is seen in the break in the line chart just before the middle of the year.

Average daily ridership is shown below by year

13_daily_max_temperature_vs_ridership.png

  1. An inflection point in maximum daily temperature was visible at approximately +15C during 2018 and at +12C during 2019. Above these temperatures, Casual ridership (dark green) grew faster as temperatures increased than it did for colder temperatures. This threshold temperature has been visually dropping in each year since 2019, suggesting increased resilience by Casual users to colder temperatures that are above the freezing point (0C).
  2. Daily ridership dependence on maximum daily temperature overlaps between Casual and Annual users. This is true at both types of bike share stations.

Observed Patterns

Daily ridership shows a strong yearly seasonality and is strongly correlated to the range of daily temperatures at both types of stations.

📈 Trend Changes

Daily ridership dependence on maximum daily temperature has shown the following changes at both types of stations

  1. Since 2018, Casual ridership has shown increased resiliency to falling daily temperatures. As a result, it has increased more sharply at a lower temperature in 2022 than it did during previous years.
  2. There is also now a strong similarity between the Annual and Casual ridership trend as a function of daily maximum temperature. In 2018 and 2019, these relationships to temperature were distinct from each other. Both trends are observed at both types of stations.

✏️ Recommendation

<aside> 💡 As suggested based on the monthly observations, the campaign should run during the warmer months of the year (prime bike share season). Casual bike share users have been dominating overall ridership and are showing an increased tolerance to colder temperatures. So, even though the peak in ridership occurs during the summer months of July and August, the campaign should prioritize running on days when the weather forecasts a maximum daily temperature that is above the freezing point (0C). This will maximize awareness among Casual members.

</aside>