Leveraging Citizen‑Generated Data in the Age of AI - and how we're making that happen

Daniel Odongo
May 16, 2024

We are at a pivotal chapter in Ushahidi's journey. Our focus is sharply on enhancing our product offerings to meet the growing demands of data management and pioneer innovative solutions that constantly address challenges in the world today. This evolution builds upon our enduring commitment to providing essential tools that empower people everywhere to easily gather data and generate insights that help tackle issues that matter most to them.

Building products that resonate with our vision while catering to the diverse needs of our users is at the heart of our strategy. At Ushahidi, we want to ensure that:-

  • People rarely heard by decision-makers share their data in ways that allow them to help each other and/or get the outcomes they want.

  • Decision-makers have data and insights that help them better support the communities they are committed to serving.

  • Community-generated datasets grow into public goods that serve as a resource to improve society.


Product Updates

Over the past few months, we've made feature enhancements to simplify data collection and analysis, ensuring that even users with limited technical skills can benefit from our platform. A philosophy of continuous improvement and innovation underpins this effort. Some of these improvements already available to all Ushahidi-hosted deployers include:

  • Increased reporting channels: Support for Mteja, Infobip, and httpSMS providers allows citizens to report incidents to a respective Ushahidi deployment via SMS. This is particularly helpful for those with limited internet access or who prefer text-based reporting.
  • Enhanced filtering and categorization: Improved filters and categories will make it easier for deployers and decision-makers to find specific information and reports within a Ushahidi deployment.
  • Larger CSV Imports: Expanding limitations on CSV import size allows deployers to upload larger datasets to their Ushahidi deployment, which is beneficial for deployers managing a high volume of reports.
  • Performance Improvements: Switching to a new database engine with attributes that support faster load times and improved performance.

To jumpstart the effective use of the Ushahidi platform—particularly for new deployers—this quarter, we shall be introducing a toolkit and a range of templates that serve as blueprints for various use cases. Whether it's crisis response, climate action, election monitoring, or human rights protection, these quick-start guides will help you manage data on your Ushahidi deployment effectively and efficiently.


Technology Partnerships

Ushahidi's continued evolution into a robust data platform opens up exciting opportunities for responsible integration of emerging technologies like Artificial Intelligence (AI) to enhance the capabilities of our platform. Specifically, our partnership with Dataminr under their inaugural AI for Good program has yielded three AI models for form assignment, geolocation, and translation which shall significantly amplify the effectiveness and accessibility of the Ushahidi platform for our global base of deployers. Below is an overview of these models that we are excited to start rolling out to citizens, deployers and decision-makers using Ushahidi:

  1. AI for Form Assignment

    From our firsthand interactions with Ushahidi deployers, we know that sorting through large amounts of data to find relevant information requires considerable time and effort, especially for organizations and community organizers with limited capacity. In critical scenarios, swift and accurate form assignment is valuable to escalate reports to appropriate team members and response partners. This is the basis behind Ushahidi’s AI for Form Assignment model, which shall automate the process of categorizing and assigning incoming reports based on content, urgency, and other predefined criteria. This model utilizes natural language processing (NLP) to analyze text data, identify key themes, and even assess the urgency of the report. For Ushahidi deployers, this will ensure that critical information reaches the right stakeholders quickly, enabling faster response times and reducing the manual overhead involved in sorting through data.

  2. AI for Geolocation

    Geolocation is a vital component in crisis mapping and response activities, helping to pinpoint the exact locations where events or issues occur. Crowdsourced mapping has become an integral part of humanitarian response despite the challenges associated with processing, validity and formatting. Our AI for Geolocation model refines data from the various data sources that Ushahidi offers (SMS, USSD, Email, Web and mobile applications), including text descriptions in reports and embedded metadata in media, to generate precise location mappings layered on top of OpenStreetMap, Leaflet and Mapbox. These improved geolocation capabilities shall increase the accuracy of data on maps and enhance responders' strategic deployment of resources.

  3. AI for Translation

    There are approximately 7,000 languages in the world. With Ushahidi's global reach, language barriers often pose significant challenges for deployment owners and managers. Ushahidi’s AI-powered translation model offers real-time, context-aware translation services to help citizens and deployers to navigate and utilize the platform in their native languages. We shall initially deploy the model for Ushahidi deployments with reports in Swahili before rolling it out to other languages in the coming months. This model shall support inclusivity, allowing more communities to participate actively and benefit from the platform. Our future perspective for this model is that reports submitted in any language shall be quickly and accurately made understandable to responders and decision-makers worldwide, which fosters more effective communication and collaboration within a Ushahidi deployment.

The above models will automate data processing, improve location accuracy, and translate reports in real time for wider accessibility. The initial rollout and release of the AI-supported functionalities for Ushahidi deployers is estimated to start in August 2024.

We're also excited to share that Ushahidi is collaborating with the Filecoin Foundation for the Decentralized Web (FFDW) in the creation of a permanent repository for citizen reports crowdsourced during elections. This permanent archive will be a valuable resource for anyone seeking to understand past elections and identify trends that can inform future democratic advancements. These datasets from Ushahidi’s election monitoring initiatives will be stored on Filecoin's decentralized network to ensure the data cannot be tampered with and for long-term accessibility to researchers, policymakers, and the public.

This initiative is a reflection of Ushahidi's continued dedication to empowering citizens with the tools they need to hold elections accountable and strengthen democratic institutions around the world. We'll be sharing more details about this initiative in the coming months, so stay tuned!

This transformation is not merely a shift in strategy but a stepping stone towards realizing our full potential as catalysts for impact through data.

Through our products, partnerships, and the collective effort of our team, we are setting the stage for truly transformative citizen science. We invite any existing Ushahidi deployers and organizations with use cases that might benefit from the AI-driven models mentioned above or the preservation of crowdsourced data to reach us using this form and a member of the Ushahidi team will get back to you. Join us as we continue to innovate and lead the way in turning data into action!