Projects /

Mapping Violence in Central Africa

CLIENT: Invisible Children & USAID
Role: SOLE Researcher & Product Designer
The Crisis Tracker is a geospatial database and reporting project that tracks armed group activity and conflict-related incidents in the remote border region encompassing northeastern Democratic Republic of Congo and eastern Central African Republic.
The Story
Policy experts & non-profits wanted to easily view and analyze incident data for the crisis in Central Africa but were limited by unclear navigation, a poor user experience throughout, and antiquated filtering capabilities.
As a result, they were reluctant to rely on the app for their job, and they often felt frustrated when attempting to use the service to pull data or gain an insight.
ROLES:
project background
I worked for Invisible Children from 2010 - 2014 as a Graphic Designer and then as Lead UX Designer. Throughout my time there I gained a knowledge of the conflict area and developed a vested, passionate interest in the ways that technology and activism could effect change in the region.

Having designed the branding for the Crisis Tracker back in 2012, I was extremely excited when, five years later, I was contacted about an RFP to redesign the web application. I jumped on the opportunity to be considered.

With the help of a project manager to coordinate meetings & line up research subjects, I owned the design process from start to finish
  1. Project scope & proposal submission
  2. User research & analysis
  3. Google analytic research & analysis
  4. Presentation of research findings
  5. Design sketching & concepting
  6. Prototyping
  7. UX Writing
  8. Visual Design
  9. Developer "handoff" & collaboration
  10. QA process
I conducted user interviews to gather qualitative data about how the app was being used by the three key personas: policy experts, employees in the field, and office-based employees.
Primary Goals
  1. Improve the information architecture of the home page & communicate the story better to users.
  2. Improve the map view & the way incidents are displayed and interact with each other.
  3. Give the daily power users better filtering tools to explore & analyze the data more effectively.
From interviews & survey results, I was able to synthesize the top concerns from active users. I did this primarily by using a digital card sorting method to cluster & identify common themes, quantify the results, and then identify opportunities that would have the greatest impact.
Mining the data for insights
It was clear from both the Google Analytic data AND the interviews, that the home page (map) should be replaced by a new & improved dashboard page - with a clear call to action to go explore the map.
What the analytics said:
  1. 91% of visitors bounced within seconds
  2. Visitors who clicked at least once, clicked the logo in the top left first.
  3. The second highest first click was on the "dashboard" link.
“It was not clear what the homepage map was - and why I was landed there. I wanted to have obvious things that I could just click on to explore.“
Problems exposed by the analytics:
  1. ❌ 91% of visitors bounced within seconds of arrival
  2. ❌ The lack of a clear call-to-action contributed to the abnormally high bounce rate.
  3. ❌ Visitors were not seeing what they expected. Most clicked on the logo thinking they were on the wrong page.
  4. ❌ Visitors second click was consistently to the dashboard. That page had very healthy bounce rates, implying that visitors discovered what they were expecting in a home page.
Solving the navigation & dashboard issues
  1. ✅ Dashboard became the homepage. The logo now links to the dashboard.
  2. ✅ A clear description of the website was added front & center
  3. ✅ An obvious call to action to explore the map was added
Problems with the map view
  1. ❌ Users could not view list of incidents while exploring the map.
  2. ❌ No way to visually differentiate between different types of incidents on the map.
  3. ❌ Incidents list didn’t interact with the map in any meaningful way.
The Solutions
  1. ✅ Move the map to the background of the page and let the incidents scroll on top.
  2. ✅ Show different incident types with different colors, and allow users to toggle on/off incident types.
  3. ✅ When a user clicks on an incident, center the map on that incident and open an info popover.
Explore the map →
Users can now combine and manage active filters with an easy-to-use filter bar with all re-designed components based on in-depth user interviews & surveys.
Consumer friendly filters. 
Powerful enough for the policy experts.
The Results
Almost a full year after this redesign was launched, I analyzed the Google Analytics data to see how it was performing. Both the avg. bounce rate & avg. session duration have shown huge improvements.
Personal Takeaways
Building an RFP-like process into any project planning This was my first time submitting to a formal RFP process. Building out the proposal was a challenge for me, as it was not something I've done before. The benefit to going through this process was that I had an extremely clear understanding of the project goals & deliverables, as well as the project timeline. I would consider using an RFP format in the future for any project as an alignment tool. View RFP Submission

Crafting a research presentation sometimes IS the point
This was the first project where I had to do a formal presentation of research findings, and I loved it. The process of crafting a presentation really helped me to internalize the insights pulled from user interviews and the analytics. If the project timeline allowed, I would aim to build this step into my future research & design workflow. View Research Presentation
Life isn't always pixel-perfect As a detail-oriented designer, it can be frustrating at times when the finished product doesn't reflect your "pixel-perfect" designs. But it's important to remember that not all projects budget in adequate space for this back and forth QA. I think I can do a better job mitigating this in the future by better communicating my expectations with the engineer(s) - in part by developing a more in-depth style & component guidelines as well as introducing some sort of design README document that lays out the areas I care the most about.

Cross-cultural user interviews Having the opportunity to conduct user interviews across different languages and continents helped improve the confidence I have in my interview skills. I've found I have a keen ability to put interviewees at ease quickly, allowing them to feel comfortable to open up about their problems & use cases in a natural way. I think this is one of my unique strengths as a designer, and I look forward to leaning more into this skillset in the future.
Coming Soon
Homeowner's Insurance
Design & Rapid Prototyping