When Orcasound users fill out the Google Form on the website (https://docs.google.com/forms/d/1oYSTa3QeAAG-G_eTxjabrXd264zVARId9tp2iBRWpFs/edit) to opt-in for notifications of Orcasound listening events at a hydrophone, they expect to be notified in real-time whenever there is “interesting” acoustic activity on any Orcasound location. Currently (mid-2023), an email is sent manually via Mailchimp when a live listening event has been detected by some combination of human listeners and a machine learning model.
During the persona development study that was conducted in Spring of 2020, it was discovered that Concerned Citizen Scientist users wanted to receive non-email notifications because they often missed the listening event because they were not alerted in real time that they had received the notification email. This finding was confirmed in an early 2023 survey of 27 web app users: in addition to more timely emails, SMS notifications were their highest priority.
To meet user needs based on that feedback, a notifications initiative was created. Some dev work focused on the back-end of a notification system was started by Dhananjay during the 2021 Google Summer of Code (see his project summary and pull requests). A UX team focused on notifications was also formed.
Since 2020, an AI known as OrcaHello is using machine learning to detect orca calls in the same live audio data streamed to human listeners via the Orcasound web app. When orca calls are detected with a mean confidence of >50% in any 60-second sample, OrcaHello automatically sends email notifications to Orcasound moderators, linking them to the OrcaHello moderator portal. Upon confirmation of each candidate detection event, the OrcaHello system sends an email to OrcaHello subscribers via SendMail (a different, much smaller test list than the MailChimp list). These automated emails trigger a manual send to the MailChimp list of Orcasound subscribers, as well, except in cases where the OrcaHello system detected weak signals or only the waning end of a bioacoustic event.
The notification system for Orcasound live-listening app users needs to be designed in such a way that users receive each type of notification they want. It can continue to be operated manually (e.g. by admins or moderators) in the short term, but in the long term could be automated through more advanced integration with both human and machine detection systems.
The user requested notification type is currently unknown- do users want desktop and/or tablet browser push notifications, mobile push notifications, SMS notifications, email notifications, or other form factor notifications like smart watches, smart speakers, etc.?
Regarding Calls To Actions (CTA), an ideal notification could have three options:
Listen live now (for users who get it in real time)
Listen to recording if the live event was missed
Listen to the best of recent events or all events
Each option within the notification could lead to a slightly different UI, all within the Orcasound web app.
Ideally, each option could eventually include a conservation call to action (CCTA) while listening to the live or recorded events. Stakeholder and discovery research should be done to explore how notifications could use CTAs to encourage conservation action from users.
If you would like to volunteer on the notifications team, please post a message in the Slack channel linked below.
Project brief: Orcasound Notifications Project Brief
Slack channel: #ux-notifications
Project folder: PROJECT: Orcasound Notifications
GitHub project: UX Work- Orcasound Notifications