RipWise

Safe Shores

Our team and I were initially given the broad instruction to uncover a problem related to one of the broad 17 UN sustainable development goals. Our team decided to focus on both the ‘good health and well-being’ and the ‘quality education’ goals from an Australian lens. Passionate about the ocean and beach safety, our team decided to venture into this problem space. Upon some cursory research, we uncovered that the most significant problem in beach safety on Australian beaches is rip-currents. We decided to delve deeper into this issue specifically.

Rips Suck

As we commenced our research, some salient statistics quickly became apparent: Rip currents are responsible for 75-90% of annual beach rescues in Australia each year (SLSC, 2021), making them the most pressing beach safety issue. While patrolled beaches provide safe swimming areas marked by red and yellow flags, approximately 96% of the Australian coastline (over 11,000 beaches) is unpatrolled (SLSC, 2021). During the 2022-23 summer, unpatrolled beaches accounted for 100% of rip-related deaths (SBS, 2023).

Despite efforts to educate the public, many beachgoers still don’t properly understand how to identify a rip. A survey conducted by Surf Life Saving Australia shows that even amongst those who are “very confident” in spotting a rip, only 54% could actually do so when tested (see figure 1). Moreover, over half of the surveyed participants were 'not very confident' in spotting a rip. Our primary research uncovered sentiments observed implicitly and shared explicitly in our research phase that “current beach signage isn’t enough.” Both patrolled and unpatrolled beaches have little to no information about rip safety. They do not address:

• What a rip is

• How to spot a rip

• How to escape a rip

Diving Deeper

So, our secondary research had uncovered a clear issue in both rip safety and educating people about rip currents. Next, we conducted primary research. Our primary research comprised of attitudinal methods: questionnaires and interviews which sought to uncover the perspectives and opinions of beachgoers. While contextual observation sought to compliment this with behavioural data.

We conducted 52 surveys, 10 interviews, and 6 contextual observations in the hopes of answering four main research questions:

1. How well do people understand beach safety cues?

2. Where do people learn beach safety?

3. What situation leads to people getting caught in rips?

4. Why do people not read signs?

We found that there was an emergent attitude of people being lazy or arrogant towards beach safety and thinking that “nothing would happen to me” as you may expect. But we also found — especially among Australians — that swimming at unpatrolled beaches often led people to dangerous situations after being caught in a rip.

Secondly, we found a significant gap in the general understanding of beach safety between people who grew up inside and outside Australia. This is shown in figure 2:

Finally, we found that those who were uninformed about beach safety are not engaging with the current beach safety signage. A large proportion of respondents mentioned it “takes too much time to read” and the current signage is not visually engaging. As such, we identified any solution must be visual and immediately engaging.

So, Who Are We Designing For?

From our research, we identified three primary personas seen below in figure 3:

This was later revised into a succinct description of a group of primary and secondary users below: 

Primary: People from non-coastal regions or diverse cultural and linguistic backgrounds with little to no beach safety knowledge.

Secondary: People familiar with basic beach safety (e.g. swimming between the flags), but lack confidence in swimming at unpatrolled beaches.

We also revised our problem statement to:

How can we improve rip current and beach safety education for people visiting unpatrolled Australian beaches; especially those from non-coastal areas or diverse language backgrounds with little beach safety knowledge?

Ideation

Next, we began the ideation process. Our first round of ideation we came up with three distinct ideas shown below:

Once we constructed a decision matrix with these ideas, RipWise (figure 6 above) came out as superior; offering a balance of engagement and accessible information.

RipWise

RipWise is a kiosk which displays and directs users to the safest place to swim at unpatrolled beaches. This concept uses a mounted camera and machine learning image segmentation to identify the safest swimming area. This idea places emphasis on communicating the danger of being caught in a rip, and addressing the danger of unpatrolled beaches. Our solution seeks to appeal to both our Primary and Secondary target audience through simple, interactive information, which is visually descriptive, but also helpful for someone who is slightly more experienced in beach safety.

We updated the UI layout to be much more visually appealing and mapped out the features in a little more detail before we commenced user testing our product. And here was the beginning of RipWise

Will This REALLY Help Swimmers?

We conducted all our user testing at a university-run user testing fair where we directed participants through activities which engaged with our design.

Methods used

• Five-second test

• Card-sort

• Think aloud

• Interviews

• Guerrilla testing.

There are three overarching usability goals which need to be met in our solution. Each member of our group was in charge of gathering data on one of these usability goals:

1) The kiosk attracts inexperienced swimmers to stop and look at it (Isak)

2) Users learn what a rip is and how to escape a rip from the design (Matt)

3) Users are able to take actionable steps to stay safe at the beach (Isaac)

Testing for Goal 1

• Determine the level of user engagement with the home screen

• Evaluate how accessible this application is for its intended users

• Assess the relevance of the content for the home screen to maximise its understandability

Testing for Goal 2

• Assess how well users interpret information on how to identify a rip current

• Identify users’ ability to understand and retain information on how to safely escape a rip current

• Test users’ understanding of general beach safety information (swell, UV, water temperature, etc.) and how these factors contribute to overall beach safety risk.

Testing for Goal 3

• Inexperienced swimmers at unpatrolled beaches are able to swim in the safest area of the beach.

• Inexperienced swimmers at unpatrolled beaches are able to escape a rip.

• Beachgoers are able to respond to relevant changing conditions and unexpected safety hazards on unpatrolled beaches.

Key Findings

The kiosk attracts inexperienced swimmers to stop and look at it

Isak’s five-second tests revealed that the kiosk’s information is “too overwhelming” for users in a rush, with users expressing confusion about where to look. Interviews and think-aloud tasks also showed that primary users, like Shweta and Jin, often needed multiple attempts to navigate the dashboard. To prevent user abandonment, the dashboard must focus on its core function: directing users to a safe place to swim.

Users learn what a rip is and how to escape a rip from the design

Matt’s testing identified that arrows and excessive text distracted from the core message of rip safety. While users reported learning more about rips, many were still uncertain about how to escape one. Shweta emphasised that rip safety information must be “front and centre” without distractions.

Users are able to take actionable steps to stay safe at the beach

Isaac’s think-aloud tasks revealed confusion over the map and virtual flags, with users uncertain about safe swimming zones and directions. Matt’s interviews echoed the need for clear instructions and better representation of safe swimming areas. Simplifying navigation and clarifying the ‘virtual flags’ concept are crucial.

Simplifying Designs

Considering these findings, our concept underwent significant change detailed below:

The final mockup can be viewed below:

Test, Test, Test

Before we began updating our mock-ups to HTML and CSS, we conducted A/B testing on our landing page and usability testing on whether users would be able to effectively navigate to the ‘safe swim zone’.

From the A/B testing, it was concluded that the map was much more intuitive, and less confusing for users.

In our second usability test, we evaluated if users could navigate to a designated "safe zone" using just a map and a live camera feed. At Gadigal Green, University of Sydney, we set up a proxy "safe swim zone," with one team member pointing their phone camera at the spot while another minimized the FaceTime feed and displayed Google Maps. Participants, including three inexperienced and two experienced swimmers, navigated a 50-meter distance unassisted, solely relying on visual cues. This validated the solution’s viability and effectiveness, demonstrating it works without written instructions, leveraging only visual elements for guidance.

Coding RipWise to Life

Main Page

On the main page we were able to integrate an interactive map with two ‘dummy’ points, and a path from the RipWise kiosk to the ‘safe swim zone’ using Leaflet. We added the ability to switch between the ‘livestream’ view and the ‘map/directions’ view. Moreover, we created a functioning QR code which then allows users to navigate to the ‘safe zone’ using their phone.

Rip Safety Page

The rip education page was uncomplicated, using straightforward flex-boxes to display the ‘what is a rip’ and ‘how to escape a rip’ sections. However, one difficult implementation which we managed to integrate was the ability for this information to be translatable. We did this by adding a function that replaces the text content for each sentence with the corresponding sentence in French. This function is called when the user switches the language. Currently, this functionality is only available in French.

Weather and Current Conditions

We implemented our solution by integrating APIs to fetch dynamic data that updates automatically. The OpenWeatherMap API provides weather codes, which are matched with a database of icons to display the current weather conditions. For UV index and wind speed, hourly data from the past and predicted 24 hours is collected, processed into arrays representing the current 24-hour period (00:00 to 23:59), and analysed to determine the maximum values. These values are then displayed as “Today’s max.” This process ensures accurate, real-time updates for both weather and environmental safety indicators, leveraging the functionality of two APIs effectively.

Live Updates

We first developed the kiosk’s HTML and CSS layout, including an accordion menu, and ensured it functioned correctly with JavaScript for opening/closing categories, adding/removing selections, and validating time inputs. Selections are dynamically encoded into a QR code. When users scan the QR code, their alert preferences and contact time are embedded into a URL, transferring the data to the mobile app without requiring personal details such as phone numbers. The app displays a popup confirming the preferences were successfully transferred before navigating to the mobile view. This adjustment addressed user concerns about privacy while maintaining seamless information transfer.

What Now?

Although RipWise is functional at an interactive level, our team is yet to implement the necessary background algorithms which would safely identify the ‘safe swim zone.’ This could be done in future versions by looking into software facilitated by TensorFlow or PyTorch.

In terms of a rollout plan, we would like to trial RipWise, on the most visited beaches on the NSW coastline. Through obtaining and analysing data from the Department of Tourism Research of the Australian Government, we propose trialling RipWise at eight of these most visited unpatrolled beaches seen below.

It’s Live!

You can check out RipWise here