Scrollytelling tutorial

From zero to your first report

Scroll down and watch how EDPA works, step by step.

01

Getting started

Setup in 5 minutes

Create a repo, install dependencies and set up issue types. The whole setup fits in a single terminal.

~/your-project
$ /edpa setup
✓ Created repository
✓ Dependencies installed
# or manually (outside Claude Code): curl -fsSL https://edpa.technomaton.com/install.sh | sh
✓ Initiative (PINK)
✓ Epic (PURPLE)
✓ Feature (BLUE)
✓ Story (GREEN)
02

Daily work

The team works

The team commits, reviews, tests. EDPA collects evidence automatically from every git log and pull request.

git log --oneline
PS
abc1234 Petr feat(S-200): implement OMOP parser
JU
def5678 Jaroslav review: approve PR #12
KN
ghi9012 Katerina test(S-200): add unit tests
MK
jkl3456 Marie docs(E-10): update acceptance criteria
03

Engine

EDPA computes

From git evidence, EDPA automatically computes Score, ratio and derived hours. No manual input.

Git Evidence
→
Score = JS × CW
→
DerivedHours = (Score/Σ) × Cap
JobSize 0
×
CW 0
=
Score 0
→
Ratio 0%
→
Hours 0h
04

Output

Report in 30 seconds

A single command generates a complete report for the iteration. All invariants are checked automatically.

~/your-project
$ /edpa close-iteration PI-2026-1.1
 
Person Cap Derived Items OK
Petr Svoboda 80h 80.0h 4 ✓
Jaroslav Urbanek 56h 56.0h 7 ✓
Marie Kralova 40h 40.0h 3 ✓
Katerina Novakova 64h 64.0h 5 ✓
Tomas Horak 80h 80.0h 6 ✓
Jan Dvorak 60h 60.0h 4 ✓
Eva Prochazkova 60h 60.0h 5 ✓
TEAM 440h 440.0h
 
✓ Invariants: ALL PASSED
05

Self-tuning

Calibration learns

With every iteration, accuracy improves. MAD (Mean Absolute Deviation) drops.

PI-1
MAD = 0.051
PI-2
MAD = 0.040 ↓ 20%
PI-3
MAD = 0.032 ↓ 37%
Karpathy loop: compare → adjust → repeat. Calibrated with a Monte Carlo simulation (68,000+ records, p<0.001).
06

Start

Start today

Run the demo, explore the dashboard and start using EDPA in your team.

~/your-project
$ /edpa setup
Dashboard demo → Methodology Back to home