In a world where technology can encourage positive social projects and their impact, Effixis, Emergent, and ArmenTeKort’s collaboration has made this vision a reality by offering a concrete example of making use of data for social progress.
This blog article highlights a unique collaboration that uses data science to address the issue of how we can build social support for equal educational opportunities. To bridge this gap, Emergent, a forward-thinking student association from KU Leuven, joins forces with ArmenTeKort, a non-profit organisation dedicated to connecting the worlds of the underprivileged and the advantaged.
This initiative is empowered by the Effixis team. We provided technical support and direction for this initiative, along with a Natural Language Processing (NLP) specialised consulting project. Together, we embark on a project that not only showcases the potential of data-driven solutions but also highlights the strength of collaborative efforts in creating meaningful change.
The Partners
1. ArmenTeKort: A Vision for a Society Free of Deprivation
Since its founding in 2014, ArmenTeKort has led social change initiatives in Flanders, Belgium, with the goal of developing structural solutions for poverty. With a community of over 3,000 citizens, ArmenTeKort bridges the gap between the underprivileged and the advantaged, striving for affordable housing, equal education opportunities, and meaningful employment. Their buddy program exemplifies a commitment to social connection and empowerment, setting the stage for the project’s objectives.
2. Emergent: Bridging Academia and Real-World Challenges
Emergent exemplifies the bridge between academic learning and practical application, focusing on the multifaceted nature of data science and artificial intelligence. Through collaboration with industry and academia, Emergent brings together students, researchers, and professionals, fostering a community dedicated to innovation and growth. Represented by a team of six motivated students, (maybe I would mention their names here?) Emergent brought fresh perspectives and energy to the ArmenTeKort challenge, underpinned by their drive to make a difference.
3. Effixis: A Data Science Consultancy Company
Effixis stands as a pillar of innovation and expertise, offering strategic guidance and technical support to businesses and governmental agencies. Through workshops and data science expertise, Effixis transforms AI aspirations into impactful realities. As a technical partner in this collaboration, Effixis played a role in guiding the group of talented students from Emergent, providing the necessary technical backbone for the project.
The Project
Objectives and Vision
The collaboration focused on exploiting data to empower individuals within ArmenTeKort’s buddy program.
The goals were ambitious yet clear: conduct sentiment analysis on life event reports, predict volunteer success in the buddy program, and forecast the outcomes of buddy duos. These reports are essentially written notes created during meetings between buddies.
In these meetings, buddies collaboratively document updates since their last gathering, capturing both positive and negative developments. For instance, a report may highlight a positive event, such as one buddy securing employment, or a negative situation, like the need for a buddy to find new housing.
The collaboration aimed to predict volunteer success in the buddy programme and the future outcomes of buddy duos by performing sentiment analysis on these life event reports. This strategic approach aimed to increase ArmenTeKort’s effectiveness and overall success rate in making meaningful connections while receiving support from the community.
Solutions
The project saw the development of several solutions:
- Sentiment Analysis Model: This model analyzes text inputs to classify sentiment, providing valuable insights into the evolution and impacts in the life of a buddy requester.
- Volunteer Prediction Model: A predictive tool designed to assess the likelihood of volunteers successfully integrating into the buddy program. With the goal to identify how they could change the volunteering training to increase the retention of volunteers.
- Buddy Duo Success Model: A model capable of evaluating the potential success or failure of buddy pairs, aiming to optimize matching processes for better outcomes and more successful duos.
Workshop: Upskilling with Cutting-Edge Machine Learning
The partnership resulted in an enriching workshop at Effixis’ Brussels office, focusing on Generative AI and its applications. The workshop covered advanced Machine Learning algorithms and their relevance to the project, offering students hands-on experience with technologies that could directly impact their work with ArmenTeKort. This session not only broadened the students’ technical knowledge but also provided insights into practical deployment strategies for the developed models.
The Impact
The collaboration between Effixis, Emergent, and ArmenTeKort has led to the following outputs:
- The students identified and defined key features that maximize the success rate of buddy duos, offering ArmenTeKort valuable insights for future pairings.
- Analysis suggested that the model could achieve better performances if the data provided contained more exploitable features. This implies a need for data that is richer in context and diversity, enabling more nuanced analyses and predictions.
- Recommendations for a revised data acquisition process were provided, suggesting the elimination of redundant features and the inclusion of previously uncaptured ones.
This partnership exemplifies the transformative power of data science in addressing social challenges. By combining Effixis’ technical support, Emergent’s talented- students, and ArmenTeKort’s social vision, the project enabled to recognize the significance of leveraging data to enhance impact, while also establishing strategies for the future utilization of data in decision-making and optimization processes.
Sources
[1] About ArmenTeKort: 10 years of action research
[2] Emergent Leuven: Consulting Projects
