Accelerating Tracker 2 Aggregate Analysis

Implementation - Implémentation Data Quality
, ,
1

This community innovation has been accepted at the 2026 DHIS2 Annual Conference as a physical poster.


Accelerating Tracker 2 Aggregate Analysis

t2a
t2a1920×1946 570 KB

Large-scale DHIS2 Tracker implementations often rely on hundreds of Program Indicators for analytics, but as complexity grows, computation can take several hours — creating critical delays for national health programs needing timely data.

IndicatorX addresses this by translating Program Indicator expressions into optimized database-level queries executed directly against existing DHIS2 analytics tables. A key innovation is separating expression translation from execution: indicators are parsed once using ANTLR4, converted into optimized SQL fragments, and stored for reuse. At runtime, these fragments are dynamically combined with org unit and period parameters to generate final queries — enabling parallel execution and dramatically faster results.

Fully compatible with the existing DHIS2 data model, IndicatorX requires no changes to core metadata structures. Initial results show processing time reduced from hours to minutes for environments with hundreds of Program Indicators, unlocking faster decision-making for health program managers across Pakistan and beyond.

Primary Author: ZUBAIR ASGHAR


Keywords:
Tracker, Tracker-to-Aggregate, Analytics Performance, Integration

t2a.pdf (681.9 KB)

2

This is a long-awaited and very useful solution, especially for countries adopting case-based surveillance and importing historical data into DHIS2 Tracker.
It would be interesting to know how the solution handles metadata mapping, duplicate detection, validation rules, and import error logs during large-scale migration.

3

Intresting! Like the Hours to Minute- very strong and apealling case.

4

Very interesting solution

5

Really valuable innovation, computation delays on large Tracker implementations are a real bottleneck, and translating Program Indicators directly to optimised SQL is an elegant fix.

6

Amazing work