The UX Development Story


The Problem

 

Dormant, unused data

Our relationships nowadays are formed and defined by interactions we have online, especially by texting. The problem this creates is that there is a wealth of information about your relationships with no way to visualize or analyze it.


The Solution

 

Peek into your texts

Textats visualizes your mobile interactions with your closest friends and family. By locally analyzing your text interactions, it reveals who your most contacted people are, shows who contacts the other more often, what the most used words are, and what time of day you text each the most. We gather and visualize the raw data into something beautiful, so that you can understand, nourish, and cherish your most beloved relationships. 


BACKGROUND

This app was the Final Project assignment of my Human Computer Interactions class with Danica Pok and Philip Bagal in March of 2015. Every two weeks, we submitted a deliverable such as Market Analysis, Storyboards, Wireframes, IRB Protocols, User Surveys & Testing, Data Evaluation, and the Final Report & Design. 

The UX Process

 

 

Market Analysis

The market of text message analysis tools is not very extensive, mostly due to the locked down nature of text messages on iPhones. However, since the Android operating system is more open to apps being able to pull data like text messages, we found one very outdated app called Relationship Analysis App.

The app claims to upload the user's text messages, process the data using their own Natural Language Processing (NLP) system, and show the user their relationship statuses with who they text. 

A few problems arise with such a system:

For one, privacy. Having all of the user's data uploaded to this unknown app's servers unencrypted is unsafe. 

Second, the NLP analysis provided is very rudimentary and unsophisticated. In testing the app's NLP strength, the app showed that the relationship between two of the group members was quite hostile based on their texts, which was completely unfounded.

Third, the design of the app looks like a basic wireframe that never made past a Low Fidelity prototype. The aesthetics did not match the current Material Design Guidelines that Google strongly recommends developers to follow. 

 

Functional & Nonfunctional Requirements

We defined the deliverables that Textats will include before designing or coding anything. We first wrote our ideas on paper, then submitted a formal proposal of the Functional and Nonfunctional requirements of our project.

 

 

Understanding the Users

By creating Personas, we could visualize the daily process and routines that our potential users are experiencing. By doing so, we understand how our app can fits into their lives. 

 
 
 

 

WIREFRAME & LO-FI PROTOTYPE

To visualize and test our ideas, we created a basic UI Wireframe, Design guide, and an interactive prototype to test our design choices with user feedback. 

 
 

 

Color Palette

The Color Palette was based off of pastel and neutral colors. Purple being the main color was chosen because we felt that most apps overused the color Blue, so we wanted a unique look. 

 
 

 

Feedback & Design 2.0

By doing preliminary tests, we recorded feedback from users about the app's appeal, interest, usability, and visual design.

By realizing that using interactive Card canvases to swipe through, we discovered that important content was hidden behind unclear taps and swipes, so we went back to the drawing board. 

After ditching the Material Design's strict guidelines, I redesigning the whole interface into a simpler, "Chat bubble" type of interactions, to reveal the details of each contact. 


 

Lessons Learned

  1. Find an problem, and work on finding an array of solutions. Don't be tied down to your first approach (Functional Fixedness problem).
  2. Test, Test, Test. It's hard to gain a new user's perspective when you've spent months working on your project (Theory of Mind problem).
  3. Don't ask a user what they want. You'll get better results by trying to implicitly understand what's not working by analyzing user behavior (Implicit vs Explicit memory).