Export¶
One command writes everything the session produced, for every tool you might open it in next.
Or click the ⇩ export tag in the header.
What lands¶
policies.scelo-export/
data.csv the cleaned dataset (what every script reads)
analysis.py pandas — the pane's analysis, restated
analysis.ipynb Jupyter notebook, with a plot cell
analysis.R base R — runs in RStudio with no packages
policies.xlsx Excel: summary · results · columns · data
policies.sce Scelo IDE project — File → Open picks up where you left
Where that directory goes depends on which terminal you are in. See Whose terminal it is in.
The scripts carry their provenance¶
Every generated script opens with the full account as comments: what was loaded,
every auto-clean step, the agent's reading, which analysis ran and why. Then it
recomputes the analysis from data.csv, on the same columns the pane used,
because the column heuristics are shared
between the pane and the generators.
So the script is not a transcript of the result. It is a program that produces the result, and you can change it.
Running the generated R against the same session:
$ Rscript analysis.R
scelo: 900 rows x 7 cols loaded
scelo → profile
...
scelo → Value by segment
n mean total share
Motor 185 470270.2 86999993 0.2173910
Marine 184 455145.7 83746811 0.2092621
Liability 190 434773.0 82606872 0.2064137
Property 178 419111.7 74601881 0.1864113
Engineering 163 443220.4 72244920 0.1805218
Against what the pane showed:
segment n mean total share
Motor 185 470,270 87.00M 21.7%
Marine 184 455,146 83.75M 20.9%
Liability 190 434,773 82.61M 20.6%
Property 178 419,112 74.60M 18.6%
Engineering 163 443,220 72.24M 18.1%
Same numbers. analysis.py produces them too.
The Excel workbook¶
Four sheets, in the order a reviewer wants them:
| sheet | holds |
|---|---|
summary |
the session: file, shape, clean steps, which analysis ran |
results |
the analysis output, the same table the HARD pane showed |
columns |
a data dictionary, one row per column, with its profile |
data |
the cleaned dataset |
The data sheet caps at 10,000 rows and says so on the summary sheet. A
120,000-row sheet of inline-string XML is a ~150MB file that helps nobody, and
data.csv always has everything.
The .sce project¶
This is the Scelo IDE's actual project format, magic scelo-project v1, the same
@scelo/core dataset shape, tested against the IDE's own parser rather than a
lookalike. Its activity log carries the load, clean, pick and run steps, so the
IDE's own export screens can replay them.
One deliberate gap
TUI runs map to the catalog's descriptive model and the runs record
stays empty. The IDE wants a numeric KPI headline for a result card, and
inventing one to fill the slot would put a fake number on it.
Opening what you exported¶
/open the export folder
/open excel the workbook, in Excel or LibreOffice
/open python analysis.py
/open notebook analysis.ipynb, in a Jupyter frontend
/open r analysis.R
/open sce the Scelo IDE project
/open csv the cleaned data
Each target goes to whatever should handle it rather than to the OS default in
every case: code -r for code files inside VS Code, a Jupyter frontend for
notebooks, because the OS opener would hand a .ipynb to a text editor, the OS
opener otherwise. On a plain terminal a .sce prefers the packaged IDE binary,
scelo-ide on PATH or SCELO_IDE_BIN, over whatever the OS associates with the
extension.
/open also works after /live with nothing exported yet, since the
live mirror's files stand in. /live then /open notebook is the whole Jupyter
flow.