PawSpot App and Wearable
January 2023 - May 2023
What is PawSpot?
Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur.
In this project, my team compared the current interface of Microsoft Academic's search dialogue with Microsoft’s in-house design of the interface.
Stakeholders
Microsoft Academic UX team
Project Sponsor
MAS Users
My Contribution
Visual Design
User Research
Testing
Project Goals
-
Identify usability issues with the current search dialogue.
-
Assess the redesign's support of exploratory search.
-
Present testing results and analytics to our sponsor, Dr. Vorvoreanu.
Secondary Research
To get a better grip on the project and what we were dealing with, my team and I researched what semantic search is. After reading various articles, we got a baseline understanding of what semantic search is: the ability of a search engine to understand the intent behind users’ search queries.
Below is a brief overview of our research findings.
Semantic vs Keyword
Many other search engines such as Google Scholar, Purdue Libraries, and many others, are based off of keyword search, but semantic search provides more refined results that match the searchers’ intent rather than just matching keywords.
Example: If you were to search “Is obesity a problem in America?”.
A Semantic Search would look for articles that have information related to Obesity in America without necessarily needing to look for keywords as it is understanding the intent the user has for the search and not just the words that are typed.
A Keyword Search would look for results that contain words like, “Obesity”, “Problem”, and “America” and offer results that only contained those keywords whether or not they are necessarily related to the searched question.
Exploratory Search
Exploratory search is the ability of a search engine to provide the user content that the user didn’t know they were looking for. As a result, the user learns new information and is able to discover information that is relevant to them.
Primary Research
Based on a meeting with our project sponsor, the team and I had a better understanding of our goals for our first round of testing. We knew that we had to focus on the search dialogue and the search suggestions. Here was our specific goals for our testing:
Learnability - Do users become proficient at the tasks we give them over time? Do the search results match users’ expectations? Do the users understand the search suggestions?
Usability - Are users able to complete tasks with minimal errors?
Exploratory Search - How well does the site support exploratory search?
When it came to creating the protocol for testing, we wanted to make sure that we created the most effective testing protocol that we could. We wanted questions that not only we believed were good for finding information, but also helped us gain more perspective on how our users interact with the search engine.
Below is a brief overview of our research findings.
Demographics
4 Female Grad Students 24 - 38 Years Old
Areas Of Importance
Icons Expectations Search Suggestions Learnability
Icons
Test participants were mostly unable to properly identify the icons, even after looking over the homepage and it’s section with icon legend.
Expectations
Participants expected different search suggestions in the search dialogue and did not think that all results were related to their search query, they also expected different results in the search results page and noted that the first few results seemed unrelated to their search query.
Search Suggestions
During testing of the current site, participants found that search results were often different than what they expected, even random at times, it was only when they got past the first few suggestions that they really felt the search result was relevant to the search and often stated that third results was exactly what they were searching for.
Learnability
A majority of test participants did not understand or recognize how MAS was building their search through their query, semantic search was unclear and often misunderstood and not learned.
These results helped inform our sketch ideations for our own version of the search dialogue. We felt that taking this feedback and applying what we learned to the search dialogue would better support MAS’s goal of supporting semantic search on through their search engine.
After gathering the results from our research and testing we all had a lot of ideas on what we wanted to improve on. Some of the main ideas we had in our sketches were:
-
Icons closer to the search bar with clear descriptions for easy understanding
-
Categorized icons to show entity properties
-
Searched words formed into sections to help user understand how searches are generated
-
Drop down/up function for changing entity
-
Label of each icon within search suggestions
Below are some of our sketches that include some of these main ideas.
Icons would be paired with the searched word in individual sections and the searcher would have the ability to change the icons to help them generate more accurate results to what they are searching for.
Icons were sectioned off into their own categories and placed directly below the search bar so that the user could see the icons when typing and avoid any confusion as to what the icons represent.
Icons were also able to be altered along with a key of what the icons represent on the right hand side of the site to prevent any confusion as to what the icons represent.
Sketches
Final Mockups
After we agreed on the final features we felt would help with some of the issues we found we started block framing a wireframe to provide a good visual of the features we wanted to include. Those features were:
-
Icon category dropdown
-
Searched word tags
-
Icon label
Below is our block frame that include these features.
After we completed our block framing to get the general idea of our functions across we then stepped into creating a high fidelity mockup of the wireframe(s).
In the first screen (below), each searched word is its own tag and it is paired with an entity that has been generated to it through the website. The suggested searches are the results of the tagged words and icons.
The searcher now has the ability to change the icon to better match what they are searching for. The drop down arrow allows the searcher to choose from the six icons, as shown below.
If the user chooses to change their icon(s) then the suggested searches below it will also change as well to match the changed icons.
The overall goal of our redesign was to help the user understand how their searches are being generated by the website along with a better way for them to narrow down their search to better match what they are looking for. Click See More to run through our wireframe.
