Session Room
Room 2 (Indigital)
Time Slot
Duration
40 min
Speaker(s)
Full name
Julien Hofer
Gender Pronouns
He/Him
Company
Department Geoinformation and Surveying,
Drupal name link
LinkedIn url
Full name
Christian Lamine
Gender Pronouns
He/Him
Company
Department Geoinformation and Surveying,
Drupal name link
Session track
Clients & Industry Experiences
Experience level
Intermediate
Digital participation platforms increasingly collect large volumes of citizen feedback — but meaningful analysis of this feedback remains a major bottleneck for practitioners. In the DIPAS Analytics project, we combine Drupal with AI-based natural language processing (NLP) and urban data in understanding and acting on participation data more efficiently and transparently. We will show our designed and implemented workflow for analyzing this data for the Free and Hanseatic City of Hamburg.
Prerequisite
- Beginners are welcome
- Basic familiarity with Drupal concepts (we will look into our custom modules).
- Interest in AI/NLP and data analytics (no advanced AI experience required).
Outline
Outline
- Problem Context: Participation Data at Scale
- DIPAS Analytics: Goal and Concept
- System Architecture
- AI Integration in Practice
- Showcase
- Lessons Learned
Learning Objectives
What You Will Learn
Attendees will gain:
1. Why AI for Participation Data
- How rapidly growing textual feedback from public participation posed a challenge.
- Why traditional manual analysis wasn’t scalable and how AI/NLP augments human review (dipas.org).
2. Technical Architecture
- How Drupal was used as the backbone for content and data management.
- How we implemented our AI services and integrated it into the Drupal World.
3. Real-World Insights
- Lessons Learnd from a 3 years project to develop our DIPAS analytics pilot
- Challenges in model selection, evaluation, and how to maintain transparency and user control.
4. Ethics & Governance
- Balancing automated analysis with human oversight.
- Approaches to maintain traceability, avoid bias, and support auditability in civic tech contexts.
5. Drupal-Friendly Takeaways
- Module and API patterns for integrating AI services in Drupal.
- Visualization and dashboard integration in Drupal sites.
