A summary of recommendations for the running the ad campaign is shown below and is based on the findings in 🔬 Exploratory Data Analysis (EDA).

This is a data-driven approach to discovering and analyzing temporal and geospatial patterns in bike share user behavior and using these insights to recommend an optimized hourly advertising preferences. These recommendations are intended to guide the marketing team’s advertising initiatives in the right direction in order to efficiently target the pool of bike share users with the right messaging and content during the digital advertising campaign.

Geospatial recommendations address the question of which stations to target. Temporal recommendations address the question of when stations should be targeted. Based on insights found during the EDA step, temporal recommendations also include the type of user to be targeted and the suggested theme of content to be shown in ads. Each type recommendation is exported to a separate file. Both files will be provided to the client.

Temporal

A recommended schedule is exported to disk with the following

1. Prioritize months between May and October of 2023, since this is prime bike share season
2. Display ads based on the following schedule

   1. weekdays (May to October)

      1. give priority to commuter-driven usage patterns
         1. 7AM to 9AM (morning)
         2. 3PM to 6PM (early evening)
      2. target
         1. Annual and Casual members
      3. theme of ad content
         1. target hybrid workers
            1. these are bike share users who use the service for commuting
            2. they follow commuter-driven temporal patterns in bike share ridership, which means they primarily use the service as part of their morning and evening commute
            3. they use the service less when the maximum daily temperature is below freezing (0C) and their ridership is at its maximum during the prime bike share season (May 1 to October 31)
            4. they are a mixture of Casual and Annual bike share members, with Casual members’ ridership dominating overall ridership across the network during 2022

   2. weekdays (November and December)

      1. give priority to leisure-oriented uses of the service
         1. 3PM to 5PM (early evening)
      2. target
         1. Annual and Casual members
      3. theme of ad content
         1. target hybrid workers (same as a., b. and d. from above, excluding the morning)

   3. weekends (May to October)

      1. 1PM to 6PM (afternoon and early evening)
      2. target
         1. Casual members
      3. theme of ad content
         1. they are leisure-oriented bike share users, who use the service for weekend activities such as running errands, meeting up with friends, going grocery shopping, etc.
         2. they have fewer constraints on patterns of usage, which means they primarily use the service during a wide window of early-afternoon to early-mid evening on weekends
         3. they also account for maximum bike share ridership during the prime bike share season, with a sharp drop-off during the colder months of the year
         4. they are primarily Casual bike share members whose ridership, as mentioned above, grew strongly during 2022 and faster than that of Annual members

      If the ratio of Casual-to-Annual member ridership returns to pre-2021 levels then weekends are not recommended for advertising.

Geospatial and Overall Performance

A list of recommended stations is exported to disk

1. Select from one of two possible geospatial groupings of the recommended stations
   1. stations located in downtown neighborhoods
      1. target
         1. these are most commuter-heavy neighborhoods in the downtown part of the city, including the financial district
      2. characteristics
         1. ridership at stations in these neighborhoods likely shows more commuter-driven patterns
         2. they primarily capture ridership by hybrid workers (who are a mix of Annual and Casual members) who use the bike share bicycles as last-mile commuter devices during the workweek
         3. they are top-performers
            1. overall
            2. on weekdays and on weekends
            3. on weekends only
   2. stations located in neighborhoods immediately to the west of downtown
      1. target
         1. these are mostly lake front neighborhoods with an abundance of recreational facilities (bike path, running trail, beach, etc.)
         2. the minority of these are stations used mostly by university students
      2. characteristics
         1. ridership at stations in these neighborhoods is likely dominated by more leisurely-oriented patterns
         2. they primarily capture Casual users who dominate bike share ridership on weekends and use the service for activities such as running errands, grocery shopping, meeting with friends, social gatherings, exercising, etc.
         3. they are stations that are top-performers
            1. overall
            2. on weekends only

Summary

Below is a summary of metrics about the recommended stations and hourly schedule to be prioritized during the campaign

The second row of this table assumes that the recommended weekend schedule is included in the campaign.

Miscellaneous

1. Toronto bike share members can check in and check out a bike from a station using an app on their phone. They would use their phone twice during each trip - to unlock (when departing) and dock (when returning) their bike. Not all members use this approach. Some use other payment methods (credit or debit card) that are accepted at most stations. In addition to displaying ads on the face of a station, to help further raise awareness among bike share users, a limited number of stickers advertising MCU’s SCS program could be printed and given to a random selection of bike share users to attach to the back of their phone and help raise awareness around the MCU program offerings among fellow bike share members.