The NG9-1-1 Address Points stage turns the address layer you already keep into the layer NENA's NG9-1-1 standard asks for: the Site/Structure Address Point layer, with every address broken into its parts the way the standard writes them.
It works for the United States and for Canada, including addresses written in French.
What the stage produces
The stage has two outputs:
- Ready: rows that meet the standard. Write them with the GeoPackage writer and you get the layer, named as the NENA template names it, with its fields in the template's order.
- Needs Review: rows the stage could not complete, each with the reason and the field it concerns. Nothing is dropped silently.
The stage does not certify that your data complies with NENA standards, and it does not submit anything to your state or provincial aggregator. You stay the authority for your addresses.
A first flow
- Add your address layer to the canvas with Dataset.
- From the Addressing category, add NG9-1-1 Address Points and connect the layer to it.
- Add a writer and connect it to the stage's Ready output. Choose GeoPackage; the layer name is filled in for you.
- Open the stage and work through its sections, top to bottom. Save and Close stays unavailable, with the reason beside the field, until the required settings are complete.
- Run the flow. The stage's node shows how many rows are Ready and how many are in Needs Review.
Standard and Address
Choose the Country. The stage then applies that country's version of the standard and its own rules: in Canada, either the county or the municipality field may be required, and the postal code takes Canada's form.
Then tell the stage about your data:
- State or Province: one value for the whole layer, or a column that holds it.
- Agency Domain: your agency's web domain, for example
harveycounty.ks.us. Only you can supply it, and Tamaat AI never suggests it. - NGUID: a column with a stable id for each address, from which the stage builds the standard's id.
- Date Updated: a column, or the time of the run.
- Address: whether the address is one column (the full address, or the street part only) or already split into parts. When it is one column, the stage reads it into its parts.
- Location: the layer's geometry, or longitude and latitude columns.
Field Mapping
One row per field of the standard, grouped as the standard groups them. Required fields come first. Show All 58 Fields lists the rest.
For each field, Value From says where its value comes from: a column, the parts read from the address, a fixed value, or nothing. Sample Result shows what the stage produces from your first rows, and Status says whether that value meets the standard.
Missing Values
Some fields are required but empty in many rows; the county or the municipality are typical. Each such field is listed under Missing Values with what is left to do.
For each field you can:
- fill it from a boundary layer: a polygon layer of your own, or a reference boundary layer, where each address takes the value of the area it falls in;
- set values area by area under Accepted Values: for example, every address in one postal community gets one county;
- ask Suggest With Tamaat AI for candidates, which you accept or dismiss one by one;
- or mark the field Left Empty when it should stay empty.
Values you accept are kept with the stage and used by every later run. A copy of the flow takes a copy of them.
Hard-to-Read Addresses
Some addresses can't be read with certainty: notes inside the address, missing separators, parts in an unusual order. They go to Needs Review.
You can leave them for review, or choose Ask Tamaat AI to Read Them. Before anything is sent you see how many addresses, what will be sent (only the address text, the city, the state or province and the postal code, never other columns), what it is expected to cost, and who processes it. Choose:
- Read Now for up to 500 addresses, ready in minutes;
- Batch for any number, at about half the cost, ready within a day. A batch is offered when your organization uses its own key with a provider that has one; on Tamaat credits, addresses are read now.
You can keep working, and Tamaat tells you in your Inbox when the readings are in. Under Readings to Review, readings are grouped by how sure Tamaat AI is. Accept all High readings in one step, accept High and Medium, or go through them one by one. Check 20 Random Readings is a quick way to judge a group before you accept it. A reading Tamaat AI could not make stays for you to decide.
Nothing Tamaat AI suggests is used until you accept it, and every accepted value is checked against the standard again on the next run.
Reading Needs Review
After a run, open the data dock and switch Output to Needs Review. Filter by Rule to work through one kind of problem at a time. Each row says the field and the reason, and its action opens the place in the stage where you fix it: map a local spelling of a street type, choose a reading for all similar addresses, fill missing values, or show the source row. CSV exports what you see.
What Tamaat AI sees, and when
Tamaat AI is used only when you ask for it: Suggest Mapping, Suggest With Tamaat AI in Missing Values, or Ask Tamaat AI to Read Them. Each time, the stage builds what is sent from named fields only, and shows it to you first. A web search runs only after you approve the exact search.
Read How Tamaat AI Handles Your Data for who processes it and under which terms.
Example: a county in Kansas
Harvey County keeps one column, FULL_ADDR, with addresses such as 123 N Main St, Newton, KS 67114, and points for each address.
- Country: United States; State: Kansas, fixed.
- Agency Domain:
harveycounty.ks.us. - Address: one column, the full address,
FULL_ADDR; Location: geometry. - Missing Values: the county is filled from the county boundary layer; the municipality from a polygon layer of incorporated places.
- Run. Most rows are Ready. A few hundred addresses with notes such as
(REAR BLDG)are read by Tamaat AI, and the High readings are accepted in one step. - The GeoPackage loads into the NENA template as it is.
Example: a rural municipality in Quebec
A municipality keeps addresses in French, such as 1250, rue de la Montagne, Saint-Éloi QC G1A 1A1.
- Country: Canada; Province: Quebec.
- Address: one column; the stage reads French street types and separators (
rue,de la) as the Canadian standard writes them. - The standard requires the county or the municipality: choose the municipality, filled from the census subdivision boundaries.
- Run, review, and write the GeoPackage.
Common Errors and Recoveries
| What you see | What to do |
|---|---|
| Save and Close is unavailable | Read the reason beside the field it names, and complete that setting. |
| Many rows need review for one field | Filter Needs Review by Rule, then fix the field once under Missing Values or Field Mapping. |
| A reading job says it cannot start | Another reading of the same stage is still running. Wait for it, or cancel it. |
| Tamaat AI is not offered | Tamaat AI must be on for your organization, with credits or your own key. See Configure an AI Provider. |