Mandates
High-level missions and visions.
About & Methodology
How the Sense-Making Dashboard works
The dashboard was created to help UNDP and Indonesian government stakeholders explore complex village development evidence in a structured way. It organizes information from source documents into a searchable and visual knowledge base, so users can inspect policy and programme levers, compare relationships, and support facilitated discussion.
Enter dashboardWhy this platform exists
Village development and poverty alleviation involve many institutions, programmes, funding mechanisms, locations, and implementation activities. Relevant evidence is often available, but it is distributed across long policy, planning, evaluation, and programme documents.
The dashboard helps address this sense-making challenge. It gives users a structured way to explore the documented policy and programme landscape, identify relationships, and prepare better questions for dialogue and decision-making.
It is especially useful when stakeholders need to compare written policy commitments with operational knowledge from the field.
What the platform analyzes
The platform works with selected source documents related to Indonesian village development. These may include planning documents, programme guidance, evaluation reports, annual reports, and other institutional materials provided for the project.
The system does not treat documents as isolated files. It extracts and organizes policy and programme elements into a connected structure that users can search, filter, and visualize.
What is a lever for change?
A lever for change is a documented item that may contribute to village development or poverty alleviation. A lever can be broad or specific.
Examples include:
- a national mandate or mission
- a strategic objective
- a policy direction
- a programme
- an activity
- a capacity-building initiative
- a physical infrastructure project
- an institutional mechanism
- a partnership or support measure
Objectives
Goals and strategic targets.
Strategies
Policy directions, priority actions, and operational approaches.
Programmes
Structured development tracks or programme families.
Activities
More specific implementation actions, projects, and operational measures.
Methodology at a glance
- 01
Curate the evidence base
Selected documents are gathered and prepared for analysis. The platform is bounded by this evidence base, which means it only shows what can be derived from the included materials.
- 02
Extract levers
The system identifies documented mandates, objectives, strategies, programmes, activities, and other relevant interventions.
- 03
Structure the knowledge graph
Extracted levers are connected to source documents, themes, organizations, administrative areas, locations, and other related levers.
- 04
Classify and group
Levers are organized using both predefined development themes and bottom-up thematic clusters derived from the content itself.
- 05
Support visual exploration
Users can inspect the same evidence through cards, tables, networks, heatmaps, maps, and detail panels.
- 06
Ground Q&A in evidence
The Q&A view retrieves relevant graph context and source-document evidence before generating a cited answer for users to inspect.
What each dashboard view is for
Cards
Use cards to browse levers as readable summaries. This is the easiest view for getting oriented.
Table
Use the table to compare many levers at once, sort by metadata, and review structured information.
Network
Use the network to inspect relationships between levers, organizations, administrative areas, and thematic clusters.
Heatmap
Use the heatmap to compare intersections across dimensions, such as tier by lever subtype or organization by organization.
Map
Use the map to inspect where documented levers are associated with locations. Missing map points should be interpreted carefully, because not every lever has usable location data.
Q&A
Use Q&A to ask open-ended questions and receive a grounded answer with source references. Treat this as a synthesis aid, not as a final decision.
How to interpret the outputs
The dashboard shows documented evidence. It does not automatically prove implementation reality.
A relationship in the dashboard means the source corpus provides evidence or a structured analytical basis for showing that connection. It should be used as a starting point for discussion and validation.
A missing relationship does not always mean that no relationship exists. It may mean the relationship is not clearly documented in the current source materials.
Use the dashboard to ask better questions
- Does this documented relationship exist in practice?
- Are these organizations actually coordinating?
- Are these activities connected to the stated policy objective?
- Are priority locations visible in the evidence?
- Which documents support this interpretation?
Responsible use and limitations
Bounded by the source documents
The dashboard can only analyze the documents currently included in the platform. It should not be treated as a complete picture of all government or village activity.
Documented reality is not field reality
The platform helps users inspect what is written in the evidence base. Stakeholders should validate whether those documented patterns reflect actual implementation.
AI answers require review
The Q&A view is grounded in retrieved evidence, but users should still read the cited references and check the source context.
Missing data should be interpreted carefully
Some levers may lack clear organization or location information. Absence in the dashboard can reflect a documentation gap, not necessarily an implementation gap.
How the platform supports workshops
The platform is designed for facilitated sense-making. It can help a workshop group move from broad questions to evidence-backed discussion.
This makes the dashboard useful as both an analytical tool and a shared discussion object.
A facilitator can
- open a pre-filtered dashboard URL
- show the relevant cards or table
- switch to the network or heatmap to inspect relationships
- use the map to discuss geography
- ask a grounded Q&A question
- invite participants to validate, challenge, or refine the pattern
Reuse and adaptation
The approach can be adapted to other policy and portfolio contexts where teams need to make sense of many documents across institutions, sectors, and administrative levels.
The reusable parts include the document-processing approach, lever extraction structure, knowledge graph model, visual workspaces, grounded Q&A pattern, and facilitation method.
Each new context would still require careful adaptation of the evidence base, terminology, policy categories, geography, organizations, and stakeholder questions.
Technical note
The platform uses a graph database to store relationships between levers, documents, organizations, locations, and themes. It also uses semantic search and evidence retrieval to support grounded Q&A. These technical layers are used to make the evidence easier to explore; they do not replace source-document review or stakeholder validation.
Explore the evidence base
Open the dashboard to search, filter, visualize, and ask grounded questions across the current source documents.