Discovering Pathways to Immigration Justice Via Data

Poesy Chen

By Poesy Chen, Bart Skorupa,  and Alma Hartman

As part of the 2023 Data to Safeguard Human Rights Accelerator cohort, Mobile Pathways’ Data Practice Accelerator partnership is focused on democratizing immigration justice via data visualization so underserved immigrants can get fair access to justice.

The United States houses the largest immigrant population globally, with more than 45 million people hailing from other countries. Yet, despite being the backbone of the U.S., immigrants face injustice due to gaps in the immigration system. With limited legal representation for immigrants, this human rights crisis affects everyone, tearing apart families and denying millions access to fair justice.

At Mobile Pathways, we aim to increase trust and access to justice for underserved immigrants. We partner with immigration-based nonprofits and utilize our online platform, “MyCamino” to deliver text updates containing reliable legal information. Serving over 1,000,000 immigrants via 150 nonprofit partnerships, we’re collecting vast data highlighting the many issues immigrants face while struggling through the maze of immigration court. 

The Challenge

While visualization and data utilization form a core pillar of Mobile Pathways' offerings to our nonprofit partners through MyCamino, we faced challenges in maximizing its impact due to a lack of technical expertise. For example, we started noticing concerning trends regarding immigration case wait times, rescheduled rates, and decisions by specific judges. As soon as we started sharing these critical findings with immigration advocates, foundations, and human rights organizations, the demand for the data grew exponentially. Still, our team needed to gain the advanced data skills necessary to efficiently communicate our findings on immigration court data.

We partnered with the Patrick J. McGovern Foundation’s Data Practice Accelerator to fully realize the potential of our evidence, which highlighted the denial of fair access to justice for immigrants, particularly asylum seekers. We hypothesized that AI tools could be of great value in aiding people going through the asylum process to make informed decisions.

However, we acknowledge that access to justice varies significantly based on multiple factors and that essential insights such as asylum grant rates by judge, court, and country of origin must be made accessible to both immigrants and legal practitioners. In order to construct a roadmap toward achieving immigration justice, we recognize the need to consider advanced techniques such as machine learning, AI, and recommendation systems.

Intending to explore the potential of AI tools in addressing the fundamental human right of fair access to justice for everyone, we embarked on this journey at a blistering pace.

Our Approach

We analyzed over 100 million rows of information about immigration court proceedings from the public Freedom of Information Act (FOIA) library. The immigration court data is stable and reliable, as it comes from the monthly extracts published by the Department of Justice.

At Mobile Pathways, our team follows a "low-code" philosophy, utilizing enterprise-level cloud technologies like Salesforce and Twilio to inform underserved immigrants of their rights through the MyCamino platform. This approach empowers us to quickly develop new capabilities for our partners without starting from scratch.

The Data Practice team advocated for a “Reduce The New” philosophy as Mobile Pathways embarked on our data-driven justice journey. To achieve data-driven justice, the Data Practice team advised exploring existing toolsets available in the Salesforce ecosystem by beginning with the native and powerful Tableau. And to think big, they also suggested incorporating AI tools using Google Cloud technologies in our tactical roadmap to realize strategic growth. 

With our new tools in hand, we debated various target audiences and decided to design it primarily for immigration attorneys. Data Practice’s advice to “Reduce The New” was again instrumental in this decision. While the toolset could also benefit immigrants and academics, we chose to focus on immigration attorneys first due to their expertise in the field and our deep experience working with this role.

To handle the challenge of a historic influx of immigration cases in court, we acted quickly. In addition, we embraced a lean data science methodology, prioritizing rapid learning through a fail-fast approach.

We created quick prototypes with actionable data, starting with a mentality of building a skateboard instead of a car. This got immigration attorneys moving faster while gaining early buy-in and feedback toward insightful data. We recognized the importance of design during these early stages, making it easy for our partners to understand the presented information. Through various iterations of Tableau dashboards, we provided partners with key trends and analyses to enable them to draw conclusions and challenge our findings.

Early Challenges

Before building any visualizations, we had to undertake a massive data exploration. Understanding the raw immigration court data file was a significant challenge. Our team is still in the discovery phase, trying to identify what data is available and what is not. We are currently using Tableau to undergo a lengthy data clean-up process to "denormalize" our data for use with different visualization and analysis tools.

To validate our understanding of the data, we continually checked the results data against published reports, making sure we were on the right track. Still, we are finding numerous challenges with available data and materials, so we are forming new partnerships with notable immigration advocates to hivemind the findings.

To maintain the momentum of our "build fast" lean approach, we shared these initial visualizations with immigration attorneys while we delved into countless tables of immigration data. Because our end-users were immigration attorneys who possess a deep knowledge of the field, they could help us determine which data points seemed accurate and which ones were not. 

The Journey Ahead

To ensure the success of our project, we recognized the importance of controlling scope and not overestimating our abilities. Like fitting food on a plate, only a certain amount can be managed at once. To improve access to justice, we partnered with the Data Practice Accelerator program, explicitly intending to utilize advanced Machine Learning and Artificial Intelligence techniques from the start. However, our initial data sets took a lot of work to denormalize, so we sought assistance to expand our capabilities.

At the same time, as our prototype dashboards got released into the wild, the growing excitement of even having the data elegantly visualized grew faster than expected. 

The Data Practice team supported this decision to right-size what we can do now while preparing for even more impact in the future.

We are proud that our "build fast" approach gained traction among immigration attorneys and led to discussions with noted immigration advocate organizations who realized the potential of our Tableau prototypes. 

While AI-generated insight would certainly enhance predictive analytics, the longer development times made it impractical to do everything at once. Furthermore, with a border crisis looming, we decided that the perfect is the enemy of the good. As such,  the Mobile Pathways team moved Machine Learning into a second phase to handle the growing adoption of our Tableau dashboards.

Left to Right: Ana Ortega-Villegas, Jeff O'Brien, and Poesy Chen

The Big Picture

What does the success of future phases for Mobile Pathways look like? Ultimately, our goal is to harness the power of immigration court data as evidence and radically realign immigration justice by:

  1. Improving efficiency:  By quickly analyzing large amounts of data, we can empower immigration lawyers with contextualized information to make informed decisions faster. 
  2. Increasing accuracy: By better understanding the micro-decisions made by immigration judges, our AI technology can reduce errors and inconsistencies that impact case outcomes (i.e., deportation).
  3. Highlighting and reducing bias: AI algorithms could minimize the impact of bias by making decisions based on objective criteria, which will hold the system accountable and make it more just.
  4. Increasing representation rates by encouraging more attorneys to take on cases, including pro bono cases.
  5. Increasing access to justice.

Our partners - a vast network of dedicated immigration attorneys - work tirelessly every day to help immigrants in the U.S. live better, more productive lives. Thanks to the Patrick J. McGovern Foundation’s Data Practice team, we will continue pursuing a vision wherein all immigrants are treated equally during legal procedures. Data will help shine a light where best intentions go astray while offering new pathways that ensure fair access to justice is a right for all, not a privilege for the few.

See this story on Medium from Patrick J McGovern