Dear all,
I've created a waterfall diagram in Power BI Desktop. It works without any issues with the Python script visualization. However, as soon as I upload the file to Power BI Services and generate a report to share across the company, the diagram disappears. In the visual center, the Python module is not displayed, and it seems that Python code cannot be used in Power BI Services. This is quite concerning, as I'm almost finished and it worked really well, but now it seems like my work has been undone. Could you please take a look and let me know if anyone has any tips?
PowerBI Desktop:
![Waterfall.png]( "Waterfall.png")
Error in PowerBI Services (Cloud):
![Error_Python.png]( "Error_Python.png")
import pandas as pd
import matplotlib.pyplot as plt
# In Power BI wird das Dataset bereits bereitgestellt. Falls nicht, wird hier ein Dummy-Dataset erstellt:
try:
dataset
except NameError:
data = {
'Application': ['App A', 'App A', 'App A', 'App A',
'App B', 'App B', 'App B', 'App B'],
'Week': ['Week 1', 'Week 2', 'Week 3', 'Week 4',
'Week 1', 'Week 2', 'Week 3', 'Week 4'],
'Costs': [100, 150, 130, 170,
200, 180, 210, 230],
'YearOnly': [2024, 2024, 2024, 2024,
2024, 2024, 2024, 2024]
}
dataset = pd.DataFrame(data)
# Füge das Jahr als eigene Spalte zur Woche hinzu
dataset["Week_Year"] = dataset["Week"] + " (" + dataset["YearOnly"].astype(str) + ")"
# Hole alle eindeutigen Anwendungen
applications = dataset['Application'].unique()
n_apps = len(applications)
# Erstelle Subplots – eine Zeile pro Anwendung
fig, axes = plt.subplots(n_apps, 1, figsize=(10, n_apps * 6), sharex=False)
if n_apps == 1:
axes = [axes] # Falls nur eine Anwendung vorliegt
# Für jede Anwendung:
for ax, app in zip(axes, applications):
# Filtere die Daten für die jeweilige Anwendung und sortiere nach Week_Year
app_data = dataset[dataset['Application'] == app].copy()
app_data = app_data.sort_values(by="Week_Year").reset_index(drop=True)
# Berechne Delta (Differenz der Kosten) – erste Woche als Startwert
app_data["Delta"] = app_data["Costs"].diff().fillna(app_data["Costs"])
# Berechne die kumulative Position (Boden der Balken)
app_data["Position"] = app_data["Delta"].cumsum().shift(1).fillna(0)
# Füge einen finalen Balken hinzu, der den Endwert anzeigt
final_row = pd.DataFrame({
"Week_Year": ["Final"],
"Costs": [app_data["Costs"].iloc[-1]],
"Delta": [app_data["Costs"].iloc[-1]],
"Position": [0] # Finaler Balken beginnt am Boden
})
app_data = pd.concat([app_data, final_row], ignore_index=True)
# Farben: Grün für positive Delta, Rot für negative, Blau für den finalen Balken
colors = ["green" if x > 0 else "red" for x in app_data["Delta"]]
colors[-1] = "blue"
# Berechne den maximalen y-Wert inkl. 20% Puffer
max_y = (app_data["Position"] + app_data["Delta"]).max() * 1.2
# Zeichne das Balkendiagramm
ax.bar(app_data["Week_Year"].astype(str), app_data["Delta"],
bottom=app_data["Position"], color=colors, edgecolor='black')
# Titel und Achsentitel setzen
ax.set_title(f"Waterfall Diagram for {app}", fontsize=14)
ax.set_xlabel("Week (Year)")
ax.set_ylabel("Costs in €")
# Werte auf den Balken anzeigen (mit Euro-Zeichen)
for i in range(len(app_data)):
ax.text(i, app_data["Position"].iloc[i] + app_data["Delta"].iloc[i] / 2,
f"€{app_data['Delta'].iloc[i]:.2f}", ha='center', va='center', fontsize=10, color='black')
# Setze die Y-Achsen-Grenze
ax.set_ylim(0, max_y)
ax.set_xticklabels(app_data["Week_Year"].astype(str), rotation=45)
plt.tight_layout()
plt.show()
