Lab: Building MMR Analytics in Superset

Register the dataset

Start by deciding the required data elements, then register them in Superset as a Dataset — a physical table or a SQL view. The dataset becomes the single source that every chart queries, which is what keeps definitions consistent across the whole dashboard.

From DHIS2, the model needs:

• Live births — the denominator
• Maternal deaths — the numerator
• Cause of death
• Facility and facility type
• Province and district
• Date of death

Then enrich the dataset with case-level MPCDSR fields: the three delays, referral status, avoidability, age, and parity. The routine data gives the rates; the surveillance fields give the diagnostic depth.

Dataset: maternal_events
Columns (12)
  live_births                    bigint
  maternal_deaths                bigint
  cause_of_death                 varchar
  facility                       varchar
  facility_type                  varchar
  province                       varchar
  district                       varchar
  date_of_death                  date
  delay_1 / delay_2 / delay_3    varchar
  referral_status                varchar
  avoidability                   varchar
  age                            integer
  parity                         integer

Estimating the MMR from routine data

The maternal mortality ratio is maternal deaths divided by live births, multiplied by 100,000. You can reach the same number two ways — prototype it in SQL Lab, or define it directly as a Superset metric on the dataset. Either way it becomes the one formula every chart on the dashboard reuses, so there’s no drift between charts.

-- Maternal Mortality Ratio (per 100,000 live births)
SELECT
  year,
  facility_type,
  COUNT(*) FILTER (WHERE event = 'maternal_death') AS deaths,
  COUNT(*) FILTER (WHERE event = 'live_birth') AS births,
  ROUND(
    COUNT(*) FILTER (WHERE event = 'maternal_death')::numeric
    / NULLIF(COUNT(*) FILTER (WHERE event = 'live_birth'), 0)
    * 100000, 0
  ) AS mmr
FROM maternal_events
GROUP BY year, facility_type
ORDER BY year;

ROUND(
  COUNT(*) FILTER (WHERE event = 'maternal_death')::numeric
  / NULLIF(COUNT(*) FILTER (WHERE event = 'live_birth'), 0)
  * 100000, 0
)

Metric: mmr (custom SQL)
Preview query result
 year | facility_type | deaths | births  | mmr
------+----------------+--------+---------+------
 2019 | tertiary       |    612 |  48,000 | 1275
 2023 | tertiary       |    354 |  48,000 |  737

Turn the metric into a chart

Choose a chart type that matches the question. A trend question calls for a line chart; a ranking question calls for a bar chart; a single headline number calls for a big-number tile. Plotting the saved MMR metric by year for tertiary facilities, and as national big-number tiles, produces the headline figures:

The tertiary-facility MMR falls steadily across the reporting period, and the national figure of 77 is tracking comfortably below the 86 ceiling set for 2025/26.

Line chart "Tertiary-facility MMR by year"
  2019 ● 1,275 per 100,000 live births
  2023 ●   737 per 100,000 live births
  trend: downward

Big Number tile "National MMR"    → 77 per 100,000 live births
Big Number tile "HSSP V ceiling"  → 86 per 100,000 live births

Assemble the dashboard

Finally, group the charts into dashboard tabs. The tabs of the Maternal Situational Analysis dashboard mirror the four analytical questions from the foundations session — from mortality overview through to circumstances of death and the care continuum.

Dashboard: Maternal Situational Analysis
Tabs (9)
  1. Maternal Mortality
  2. Overview of Deliveries
  3. Regional Investigation
  4. Maternal Death Audit
  5. Mother Characteristics
  6. Circumstances of Death
  7. Facility Analysis
  8. Surveillance
  9. ANC and PNC