medfret_OUT.xlsx, live formulas — see link top-right) exists alongside this dashboard. The two are fully independent: this page never reads the workbook, and the workbook is never regenerated from this page. Updating one does not update the other — regenerate the workbook manually (python3 build_medfret_excel.py) if you want it to reflect a change made here.
Cleaning Method: IQR × k
Calls outside Q1−k×IQR / Q3+k×IQR excluded. Negative and missing values also removed.
Scatter: All Port Calls by Volume Class
Fort — Reference Table
Moyen — Reference Table
Faible — Reference Table
All Port Calls (cleaned)
| # | ID | Volume Class | Duration (days) | Status |
|---|
Histogram (cleaned data)
Cumulative Distribution — CDF
Model Fit Comparison — CDF MSE — 6 models
Optimal Model Parameters
Validation Statistics
Fort — KPI Status
Moyen — KPI Status
Faible — KPI Status
KPI Summary Table
| Volume Class | Target (days) | DB Avg | DB Std | CV | Margin | P(success) |
|---|---|---|---|---|---|---|
| Fort | 3.0 | 2.66 | 0.46 | 17.3% | ±0.19d | 77% |
| Moyen | 2.0 | 1.58 | 0.53 | 33.7% | ±0.12d | 78% |
| Faible | 1.0 | 0.81 | 0.33 | 40.0% | ±0.04d | 72% |
| Overall | — | — | — | — | — | 72% |
Overall = min of all three. Faible is binding at 72%. Need target ≥ 0.94d for 80% on Faible.
Target Sensitivity — Faible Class
Fort
Moyen
Faible
Fort — MC Distribution
Moyen — MC Distribution
Faible — MC Distribution
Monte Carlo Summary Table
| Volume Class | DB Mean | P20 (conservative) | P50 (median) | P80 (optimistic) | Spread P80−P20 |
|---|---|---|---|---|---|
| Fort | 2.66 d | ~2.27 d | ~2.67 d | ~3.08 d | 0.81 d |
| Moyen | 1.58 d | ~1.11 d | ~1.58 d | ~2.01 d | 0.90 d |
| Faible | 0.81 d | ~0.53 d | ~0.81 d | ~1.09 d | 0.56 d |
Decision Summary — Include in Rotation?
| Volume | Mean | Target | P(success) | Decision |
|---|---|---|---|---|
| Fort | 2.66d | 3.0d | 77% | Include |
| Moyen | 1.58d | 2.0d | 78% | Include |
| Faible | 0.81d | 1.0d | 72% | Include / Review target |
Best Statistical Model per Class
MC P20/P50/P80 Summary
Comparison with Prévision projet Methodology (R&D Projects)
| Metric | Prévision projet (R&D Projects) | MEDFRET (Port Calls) | Note |
|---|---|---|---|
| Variable | Duration (months) | Duration (days) | Same concept, different scale |
| Classes | 1st in Fam / Innovative / Follower | Fort / Moyen / Faible | 3 complexity/volume levels |
| Cleaning | IQR×2.5 | IQR×2.5 | Identical method |
| Outliers removed | 2 / 151 = 1.3% | 15 / 342 = 4.4% | Higher data quality issues in logistics |
| Best model (high) | Log-Normal | Log-Normal | Consistent right-skew behaviour |
| Best model (low) | Log-Normal / Weibull | Gamma | Faible more symmetric |
| Monte Carlo | Box-Muller N(μ,σ) 1,000 trials | Box-Muller N(μ,σ) 1,000 trials | Identical method |
| KPI probability | 29% (duration) | 72–78% | Port targets more conservative relative to mean |
| Pearson r (QQ) | 0.988–0.998 | 0.955–0.978 | Good but slightly lower — logistics more noisy |
By installing, accessing, or using this Software, you ("User") agree to be bound by this End User License Agreement ("Agreement"). If you do not agree to these terms, do not install, access, or use the Software.
Subject to the terms of this Agreement, the company ThinknSolv grants the User a non-exclusive, non-transferable, revocable license to use the Software for personal or business purposes, strictly in accordance with this Agreement.
The User may NOT copy, modify, distribute, sell, or lease any part of the Software, nor transfer the license to another party without prior written permission from ThinknSolv.
All intellectual property rights in and to the Software, including but not limited to copyrights, trademarks, and patents, remain the sole property of ThinknSolv. This Agreement does not transfer any ownership rights.
This license is effective until terminated. ThinknSolv may terminate this Agreement if the User fails to comply with its terms. Upon termination, the User must cease using the Software and delete all copies.
The Software is provided "AS IS," without warranties of any kind, express or implied, including but not limited to merchantability, fitness for a particular purpose, or non-infringement.
To the maximum extent permitted by law, ThinknSolv shall not be liable for any direct, indirect, incidental, or consequential damages arising from the use or inability to use the Software.
This Agreement shall be governed by the laws of France. Any disputes arising under this Agreement shall be resolved through arbitration in Marseille or in a court of competent jurisdiction.