pima
pima

/ˈpi.ma/

Swahili · verb

to measure

A platform for the people who  have the data but not the answers.

pima reads the systems you already run and turns them into numbers you can act on.

Request evaluation access How it works

What pima is

A domain-aware analytics and reporting platform. Every figure is defined once, and dashboards, statutory reports, exports and the AI assistant all answer from that one definition.

pima is deployed into the environment that already holds your data and reads it, read-only. It builds one conformed warehouse from the operational systems you run, defines every figure in a metric catalog, and decides who may see what once, before anything is shown. The subject-matter knowledge comes from a domain package on top of a core that knows no domain.

How it works

Five steps between your source system and the people who read the numbers.

  1. read An adapter opens the source read-only. Nothing writes back.
  2. derive A warehouse is rebuilt from the source on a schedule, in the shape the facts take.
  3. define One definition per figure, in a catalog every output reads.
  4. decide Access is settled per person, at one point, before anything is shown.
  5. serve Dashboards, statutory reports, exports and the AI assistant, all agreeing.

The five-minute version

Two domains

Each is a package of subject-matter depth on the same platform. A deployment serves one and is built for it.

A patient-flow view in pima's health domain: people starting treatment, retained, interrupting and returning, over time.

Health

HIV from testing through treatment to preventing transmission from mother to child. Tuberculosis, drug-susceptible and drug-resistant. Maternal and child health for every pregnant woman and her child. Cascades that end in the people behind the drop, and the returns due each quarter.

The health domain
A portfolio-at-risk view in pima's financial-services domain: PAR 30, 60 and 90 by office and product.

Financial services

A lender's loan book read from the core banking system as it changes. Portfolio, risk and financial performance, each figure carrying the date it is true for, fresh as the branch. Lending first, starting with the microfinance institution.

The financial-services domain

Four things worth knowing early

Nothing writes back.
pima reads your source systems and never modifies them. A data-quality problem is surfaced, not quietly corrected.
Every number can be traced.
Any figure on a dashboard or in a report follows back through its transformations to where it came from.
One definition, many outputs.
A dashboard, a report, an export and an AI answer read the same definition, so they agree.
Access is a decision, not a screen.
Who can see identifying data is settled in one place and applied everywhere. Redaction is always visible, so missing and restricted are never confused.

See it working

Two evaluation instances run on generated data, one for each domain. Twenty minutes on one settles what a description cannot.

Access is granted on request. Every request is read by a person, and you can ask for either instance or both.

Request evaluation access What an evaluation account sees