Download your data with its labels
Celeus keeps a codebook with each dataset: a label, a unit and a description for every column, value labels such as 0 = bad and 1 = notbad, and the codes that mean missing. Download data writes a dataset version together with that codebook in a standard format, so the labels survive the trip into other software. There is no Celeus-specific file format.
Download a version
Section titled “Download a version”- Open the dataset.
- Open the More menu at the top of the page and choose Download data.
- Pick a format and a version, then choose Download.
The file holds every column of the chosen version, including columns you excluded from the AI, with the values exactly as they are stored. Downloading is free and does not use an analysis or any AI.
Which format keeps what
Section titled “Which format keeps what”| Format | Labels and missing codes |
|---|---|
SPSS (.sav) |
Variable labels, value labels and user-missing codes. Up to three missing codes are marked missing; a range is used when it contains no real value. |
Stata (.dta) |
Variable labels and value labels on whole-number columns. Stata has no user-missing values for arbitrary codes, so missing codes stay as values with a “(missing)” value label. |
Excel (.xlsx) |
A data sheet and a codebook sheet. The codebook sheet can be uploaded back as a codebook. |
CSV with codebook (.zip) |
The data as UTF-8 CSV, the codebook as CSV and as JSON, and a notes file when something could not be carried over. |
Data Package (.zip) |
An open Data Package with a table schema, value categories and missing values. |
Column names that a format does not allow (SPSS and Stata restrict length and characters) are changed to a valid name, and the original name becomes the variable label when the column has no label of its own. The notes explain every change.
Text that starts with = + - or @
Section titled “Text that starts with = + - or @”Spreadsheet programs can run text that starts with =, +, - or @ as a formula. In the Excel file, text cells that start that way get a leading apostrophe so they open as text, and the notes say how many cells in which column were changed.
In the Excel, CSV and Data Package files, a column name that starts that way gets the same apostrophe, and the notes say so; remove the apostrophe to match it by name, because the JSON codebook keeps the original name.
CSV and Data Package cell values are written exactly as stored, so a text value such as -9 still matches the missing code declared in the JSON codebook and in the Data Package schema. The CSV codebook keeps its apostrophe guard (it shows '-9, and Celeus removes the apostrophe when you import it). Open the data files as text, or import them with the column types set. The SPSS, Stata and JSON files are not changed.
De-identification and who can download
Section titled “De-identification and who can download”Each version in the list shows whether it is de-identified, contains identifiers, had identifiers detected and reviewed, or had identifiers detected that were never enforced or reviewed. A version with detected identifiers can be downloaded only when identified storage is enabled for your workspace and the identifiers were enforced or reviewed; otherwise choose a de-identified version. Every download is recorded in the audit log with the dataset, version and format, never with the dataset name.
People who can view a dataset in your workspace can download it. Researchers who only have access through a shared project cannot.