What we build

A rate consultant, an analyst pricing a credit and a utility finance director need three different things from the same record. These are the pieces; each page says what it is built from and what it does not do.

Rate models and exhibits

The calculation layer of a study, built against your own method.

• Affordability exhibits — Household income and demographics joined to a service area, from the published census tables, with the join written down so the exhibit can be defended rather than only presented.
• Rate calculation modules — Cost allocation and rate design as a spreadsheet or a Python module, built against your own class structure and your own cost pools, so the model is yours to run and to defend.

Credit and transaction work

Operating condition and disclosure timing, from what a system filed with its regulator.

• Covenant and capital plan testing — A capital plan run against a rate covenant and its coverage requirement, from the audited statements, so the year coverage breaks is a date rather than a worry.
• Disclosure recency — How old the last regulator filing is for a system, against how old its last audited statement is. The gap is operating record that exists and is not yet in the disclosure.
• Lapsed and unstable filers — Systems that stopped filing under an annual requirement, and systems whose reported figures move too much to be relied on. Both are already in the record; neither is published as a list you can screen against.
• Operating condition screens — Filed loss figures, the share of each that is a regulator default rather than a measurement, and which filings move too much to rely on. The operating record a system reports, not the description in its offering document.

Pipeline automation

The assembly that happens before the work starts, done once and delivered on a schedule.

• Filing extraction — The published reporting tables pulled, checked against the agency's own arithmetic, and delivered as files on a schedule. Every figure carries the file and the cell it was read from.
• Normalized consumption panels — A utility's billing extract turned into one row per account and period, on a consistent class and meter-size vocabulary, with every period that could not be relied on flagged and the reason for the flag kept beside it.
• Peer sets and benchmarks — Normalized multi-year filings for comparable systems, with unstable and refused filings flagged rather than dropped. Delivered as a file, on a schedule, with every figure keyed to the cell it was read from.

Programme oversight

A view of a state programme built from the filings it has already published.

• Consolidation candidate screening — Which systems in a program are small, are struggling to keep filing, or are reporting figures that rest on the regulator's own default values, and which of them sit in the same county as a larger system that is not.

Who it is for

One page per kind of reader, with the pieces tagged to them.

• Cities and utilities — What your own filed record says about your system, a process for correcting it, and the studies built from your audited statements and billing records.
• Consultants — The assembly layer of a study — peer sets, models, extraction, affordability exhibits — built once against your method and delivered under your own name.
• Credit analysts — Operating condition and disclosure timing for a water or sewer system, from the filings its regulator holds rather than from the offering document.
• State agencies — A view of your own program built from your own published data: who has lapsed, and which filings rely on the defaults your form supplies.
• Vendors — Which systems have the problem your product solves, from the published record: unexplained losses, lapsed filers, and figures too unstable to act on.