Skip to content
VestaDocs home

Preparing data

Get a dataset ready for analysis: keep join keys as text, rename and remove variables, and run Change of scale and Join layers in Preprocess.

Online8 min readUpdated Nov 2026

Most of the work of preparing data happens before it reaches Vesta Online, and a little happens after. This article covers both: what your file needs before import, and what you can do to a dataset that is already in a project (rename a variable, remove one, and run Preprocess steps). The walk uses a copy of the Diabetes and poverty sample, which holds two datasets, US counties and County poverty (ACS), each with 3,109 counties. Every step below was run on Vesta Online, and the screenshots come from that run. For the import steps themselves, see Importing data.

Before you import

Section titled “Before you import”

One thing has to be right in your file, because the import wizard will not fix it for you.

Rates can wait until after import. If your file has a count and a population, Preprocess can compute the rate and smooth it: see Compute and smooth rates after import in Importing data. If your file already has a column that holds a rate, that is fine. The Variables step can mark that column’s role as Rate.

Keep join keys as text. A code such as a county FIPS is a fixed-width digit code, so it types as String. As a number, 06037 becomes 6037 and matches nothing. When you join a table to geography, the import wizard says this itself and locks the key column to String, with an explicit Unlock.

Prepare a dataset in a project

Section titled “Prepare a dataset in a project”
  1. Open the Data panelOpen the project and choose the Data tab on the right. Each dataset is a row; expand US counties to list its variables, with an icon for each variable's type.
  2. Rename a variableSelect a variable, press F2, type the new name and press Enter. Here Total Pop 18+ becomes Adults 18 and over. Vesta shows a confirmation, Renamed to Adults 18 and over, with an Undo button, and the new name is still there when you reload the project.
  3. Remove a variableSelect a variable and choose Remove at the bottom of the card. A dialog asks Remove State? and says that nothing else is removed and no panels close. Choose Remove to confirm, or Cancel to keep it. The variable leaves the list.
  4. Choose Change of scaleChoose Preprocess in the left rail, then pick the dataset under Change dataset (here County poverty (ACS)). Preprocess opens at Transformations, a plan of the steps to run in order. Clear the plan if it holds steps you do not want, then choose Change of scale, which aggregates or interpolates values from one geography onto another.
  5. Set up the stepChoose Next: Configure Change of scale. Under Destination geography keep In this project and pick US counties. Untick the variables you do not need (here GEOID and NAME) and choose a method for the one you keep: Apportion, Mean or Majority. This walk uses Mean on All people (percent).
  6. Review the planChoose Next: Review. Review lists the step with its variable and states that nothing is written until you run.
  7. Run it and read the SummaryChoose Run preprocessing. The run is a server job with a log; when it finishes the Summary says how many steps were applied and records them as provenance. The Next panel offers Done, Visualize, Analyze and Preprocess, and Remove results, which restores the project version the run replaced.
  8. Find the new variableChoose Done and expand US counties in the Data panel. The aggregated values are a new variable on US counties, here All people (percent), a Double, with State already gone from the earlier step.
  9. Choose Join layersBack in Preprocess, with US counties as the dataset, clear the plan and choose Join layers. After Next: Configure Join layers, pick the Source layer (here County poverty (ACS)) and a key on each side (GEOID and GEOID). The zero-pad option rewrites numeric keys, for example 5095 to 05095; leave it unticked for keys that are already text. Unmatched rows are kept (left outer) or dropped (matched rows only).
  10. Review the joinChoose Next: Review. This run stopped here, without running the join.

What you should see

Section titled “What you should see”

In the Data panel

Section titled “In the Data panel”
Data panelThe Data panel with US counties expanded, listing GEOID, State, County Name, Total Population, Total Pop 18+, Diagnosed diabetes and Poverty rate
US counties expanded to its variables. The card at the bottom shows the selected item and its actions: Folder, Visualize, Analyze, Preprocess, Export and Remove.
RenameThe Data panel after a reload, with Adults 18 and over in the list in place of Total Pop 18+
The project was reloaded after the rename: the new name is in the list where the old one was.
RemoveThe Remove State dialog with Cancel and Remove buttons
Remove asks first. The dialog names the variable and says nothing else is removed.
After RemoveThe Data panel after State was removed, with the list one row shorter
State is gone from the list. The Undo button at the top of the workspace is still available.

Change of scale

Section titled “Change of scale”
TransformationsThe Transformations step with Change of scale selected and listed in the plan
The plan on the right lists the steps that will run, top to bottom. Each step runs on the result of the one before.
ConfigureThe Configure Change of scale step with US counties as the destination and Mean chosen for All people (percent)
Destination US counties; GEOID and NAME unticked; Mean chosen for All people (percent).
ReviewThe Review step with one Change of scale step listed
Nothing is written until Run preprocessing.
SummaryThe Summary of a finished Preprocess run, with the run log, a result panel and the Done button
The run log lists the stages, and the Next panel offers Done, Visualize, Analyze, Preprocess and Remove results.
New variableThe Data panel with US counties expanded, listing All people (percent) as its last variable
The result is a variable on US counties: All people (percent), a Double.

Join layers

Section titled “Join layers”
Join layersThe Configure Join layers step with County poverty (ACS) as the source and GEOID chosen on both sides
Source layer, a key on each side, the zero-pad option and the choice of keeping or dropping unmatched rows.
Review the joinThe Review step with one Join layers step listed with its key GEOID
Review lists the Join layers step and its key, GEOID.

Two ways to put a table’s values on a geography

Section titled “Two ways to put a table’s values on a geography”

Change of scale moves values from one geography onto another, and Join layers attaches a table to a layer by a shared key. They overlap when two datasets share the same geometry, as the two datasets here do. Use Change of scale when the geographies differ, and Join layers when you have a table keyed by an identifier. The walk ran Change of scale only; a Join layers run would add a second copy of the poverty column, so it stopped at Review.

What Preprocess does to your data

Section titled “What Preprocess does to your data”

For the steps walked here, Review says: “Every result is written as a new variable or a new dataset. The original is left as it is.” A step that replaces its original dataset shows a different note, “One step replaces the original dataset. That cannot be undone once it has run.”, and a Not undoable badge. Change of scale, as walked here, showed the first note. Removing a variable is different: the variable leaves the dataset, and Undo is the way back.

Where next

Section titled “Where next”
Was this page helpful?
© 2026 BioMedware, Inc. · docs.vestastats.comWebsiteAcademyDesktop docsPrivacyAccessibility