LS 2.0 Facility Dashboard
Bi-weekly PHC monitoring · Kaduna State
1
6
0
100

Readiness score trend by round

Composite of six pillars: open on arrival, attendance, stock availability, vaccine availability, session completion, cold chain. Hover for the pillars.

Readiness band distribution

Share of filtered visits in each band.

Readiness by LGA

Mean composite score; dotted line is the filtered average.

Facility map

Colour = readiness, size = stock-out rate. Hover a facility for its profile.

Readiness pillars by facility type

Each pillar is a 0-100% rate; a differentiated support package by level of care is warranted when gaps differ by type.

Open on arrival, by LGA

Share of visits where the facility was already open when the data champion arrived.

Why facilities were closed

Reasons recorded when the facility was closed on arrival.

Hours of operation

By facility type (unique facilities).

Emergency referral capacity

Can refer obstetric emergencies; transport available when needed in last 2 weeks.

Cold chain equipment

Availability vs functionality where available.

Service availability & session completion

Offered = facility provides the service; completion = all planned sessions last week conducted.

Why planned sessions were missed

Service

Training coverage (past 2 years)

Share of filtered facilities with at least one staff member trained, by training type.

Attendance of scheduled permanent staff, by cadre

Present ÷ scheduled (morning + afternoon + evening shifts).

Attendance by LGA

Why permanent staff were absent

CadreTop10

Workforce composition by cadre

Permanent, ad hoc/seconded (A: BHCPF/IMPACT-funded; B: NYSC/students) and volunteers.

Salary timeliness & attendance by round

Bi-weekly 3 (round 4) shows what happens to attendance when salaries are late.

Attendance by salary timeliness and roster practice

How often staff leave the post to access salary

Attendance vs distance to LGA headquarters

Each point is a facility (mean over filtered rounds).

Stock-out rate by commodity

Category

Stock-out heatmap: LGA × commodity

Darker = higher share of visits with a stock-out since the previous visit.

Main reasons for stock-outs

Commodity

Requisition funnel

Submitted → complete & on time → fully received → delivery documented.

Stock-out rate by supplier

Requisition behaviour vs stock-outs

Requisition submitted × knows minimum-stock calculation.

Below minimum stock (early warning)

Show items ≥0%

Stock-out trend by category and round

Vaccine availability (facilities that stock vaccines)

Doses used, by vaccine

Vaccine stocking status

Unique facilities in the filter.

Availability by cold chain status

Cold chain interruption reasons

Vaccine availability by LGA and round

Models are trained on what was observed at one visit to predict the next one, so they can be re-run after each bi-weekly round. Scores below are for the latest round of each facility in the current LGA / type filter (round sliders do not apply here).

Facilities at risk of dropping below readiness 65 at the next visit

Risk threshold50%

Commodity stock-out risk at next visit

Min risk60%

Facility × commodity pairs above the threshold, ranked. Use for targeted emergency resupply.

Expected attendance gain if salary is paid on time

Counterfactual from the attendance driver model, per facility (latest round). Facilities with late salary show the gain; on-time facilities show zero.

Stock-out model: what matters

Stock-out model: ROC (held-out round)

At-risk model: what matters

Attendance driver model: what matters

Facility segments

K-means on the mean readiness pillars per facility; use segments to tailor the support package.

Thirty decision questions answered with SQL against the survey warehouse (full network, all rounds). Each links to its standalone interactive figure. The tabs above let you re-cut the same metrics by LGA, facility type and round.
The full analysis as an executed Jupyter notebook: the thirty questions (SQL → table → figure → answer), the predictive models trained and evaluated step by step, and an in-notebook filterable dashboard. It is rendered below; use the links to open it on GitHub or nbviewer, download it, or run it yourself.
Filtered visit-level records ( rows). Click a header to sort; use Export CSV above to download the current selection.