diff --git a/app/__init__.py b/app/__init__.py
new file mode 100644
index 0000000..723c323
--- /dev/null
+++ b/app/__init__.py
@@ -0,0 +1 @@
+# Package-Initialisierung für m365-workingwith
diff --git a/app/aggregator.py b/app/aggregator.py
new file mode 100644
index 0000000..3a825ea
--- /dev/null
+++ b/app/aggregator.py
@@ -0,0 +1,28 @@
+from __future__ import annotations
+from collections import Counter
+from dataclasses import dataclass
+from typing import List
+from .models import WorkingWithRelation
+
+
+@dataclass
+class DepartmentLink:
+    source_department: str
+    destination_department: str
+    weight: int
+
+
+def aggregate_department_links(relations: List[WorkingWithRelation]) -> List[DepartmentLink]:
+    counter: Counter[tuple[str, str]] = Counter()
+    for rel in relations:
+        src = (rel.source.department or "").strip()
+        dst = (rel.destination.department or "").strip()
+        if not src or not dst:
+            continue
+        counter[(src, dst)] += 1
+
+    links: List[DepartmentLink] = [
+        DepartmentLink(source_department=s, destination_department=d, weight=w)
+        for (s, d), w in counter.items()
+    ]
+    return sorted(links, key=lambda l: l.weight, reverse=True)
diff --git a/app/auth.py b/app/auth.py
new file mode 100644
index 0000000..ffaad3c
--- /dev/null
+++ b/app/auth.py
@@ -0,0 +1,51 @@
+import asyncio
+import json
+from pathlib import Path
+from playwright.async_api import async_playwright
+
+AUTH_STATE_FILE = Path("auth_state.json")
+
+
+async def ensure_login(email: str) -> None:
+    """
+    Öffnet den Browser mit/ohne bestehenden auth_state.
+    Beim ersten Mal: manueller Login in M365.
+    Danach wird der Auth-State persistiert.
+    """
+    async with async_playwright() as p:
+        browser_type = p.chromium
+
+        if AUTH_STATE_FILE.exists():
+            context = await browser_type.launch_persistent_context(
+                user_data_dir="user_data",
+                headless=False,
+                channel="chrome",
+            )
+        else:
+            context = await browser_type.launch_persistent_context(
+                user_data_dir="user_data",
+                headless=False,
+                channel="chrome",
+            )
+
+        page = await context.new_page()
+        await page.goto("https://m365.cloud.microsoft/search/?auth=2")
+
+        print(f"Bitte mit {email} in Microsoft 365 anmelden.")
+        print("Nach erfolgreichem Login Browser-Fenster schliessen oder Skript abbrechen (Ctrl+C).")
+
+        # Langes Warten; in echter Implementierung könnte man Events verwenden
+        await page.wait_for_load_state("domcontentloaded")
+        'page.wait_for_timeout(1000 * 1000)'
+
+
+
+        state = await context.storage_state()
+        # storage_state() liefert dict -> JSON speichern
+        AUTH_STATE_FILE.write_text(json.dumps(state), encoding="utf-8")
+
+        await context.close()
+
+
+def ensure_login_sync(email: str) -> None:
+    asyncio.run(ensure_login(email))
diff --git a/app/easyvisualize.py b/app/easyvisualize.py
new file mode 100644
index 0000000..4a16980
--- /dev/null
+++ b/app/easyvisualize.py
@@ -0,0 +1,101 @@
+import pandas as pd
+import plotly.graph_objects as go
+import sys
+import os 
+
+
+# --- KONFIGURATION & DATEIPFAD ---
+
+# Prüfen, ob ein Pfad als Argument übergeben wurde, sonst Default nutzen
+
+if len(sys.argv) > 1:
+  csv_dateipfad = sys.argv[1] 
+else:
+  csv_dateipfad = '/output/relations.csv'
+
+# Sicherstellen, dass die Datei existiert
+
+if not os.path.exists(csv_dateipfad):
+  print(f"Fehler: Die Datei '{csv_dateipfad}'wurde nicht gefunden.")
+
+  sys.exit(1)
+
+
+# --- DATEN LADEN ---
+try:
+
+  # Wir laden die CSV (Trennzeichen ; ist im deutschen Excel-Raum Standard)
+  df =pd.read_csv(csv_dateipfad, sep=None, engine='python')
+  print(f"Daten erfolgreich geladen: {len(df)} Zeile aus '{csv_dateipfad}'.")
+
+except Exception as e:
+  print(f"Fehler beim Lesen der CSV: {e}")
+  sys.exit(1)
+
+
+# --- DATEN AGGREGIEREN ---
+# Wir zählen die Kommunikationspfade zwischen den Abteilungen
+
+''' df_agg = df.groupby(['source_department', 'destination_department']).size().reset_index(name='weight') '''
+df_agg = df.groupby(['source_department', 'destination_displayname']).size().reset_index(name='weight')
+
+# Liste aller Departments für die Knoten-Beschriftung
+''' all_nodes = list(pd.concat([df_agg['source_department'], df_agg['destination_department']]).unique()) '''
+all_nodes = list(pd.concat([df_agg['source_department'], df_agg['destination_displayname']]).unique())
+
+node_map = {name: i for i, name in enumerate(all_nodes)}
+
+# Mapping auf Indizes
+source_indices = df_agg['source_department'].map(node_map)
+target_indices  = df_agg['destination_displayname'].map(node_map)
+''' target_indices  = df_agg['destination_department'].map(node_map) '''
+
+weights = df_agg['weight']
+
+
+# --- FARBGESTALTUNG ---
+# Wir färben die "Source"-Abteilungen (dein Team) anders ein als die "Ziele"
+node_colors = ["#1f77b4" 
+if node in df['source_department'].unique() else "#9467bd" for node in all_nodes]
+
+# --- VISUALISIERUNG ---
+fig = go.Figure(data=[go.Sankey(
+  node=dict(
+  pad=20,
+  thickness=30,
+  line=dict(color="black", width=0.5),
+  label=all_nodes,
+  color=node_colors
+ ),
+
+  link=dict(
+  source=source_indices,
+  target=target_indices,
+  value=weights,
+  color="rgba(200, 200, 200, 0.5)"
+  # Transparente graue Pfade
+
+ )
+
+)])
+
+
+
+fig.update_layout(
+  title_text=f"Organisatorische Netzwerkanalyse: Abteilungs-Flüsse<br><sup>Quelle: {csv_dateipfad}</sup>",
+  font_size=12,
+  height=800
+)
+
+# --- AUSGABE ---
+output_filename = "netzwerk_analyse_lokal.html"
+
+fig.write_html(output_filename)
+
+print(f"Analyse abgeschlossen. Interaktives Diagramm gespeichert unter: {output_filename}")
+
+# Automatisches Öffnen im Standardbrowser
+import webbrowser
+
+webbrowser.open('file://'+ os.path.realpath(output_filename))
+
diff --git a/app/extractor.py b/app/extractor.py
new file mode 100644
index 0000000..80fe9ed
--- /dev/null
+++ b/app/extractor.py
@@ -0,0 +1,174 @@
+from __future__ import annotations
+import asyncio
+import csv
+from pathlib import Path
+from typing import Any, Dict, List
+
+from playwright.async_api import async_playwright, Page, Response
+
+from .auth import AUTH_STATE_FILE
+from .parser import (
+    parse_person_from_organization_person,
+    parse_direct_emails_from_organization,
+    parse_workingwith_entries,
+)
+from .models import Person, WorkingWithRelation
+
+
+DELV_PERSON_URL = "https://eur.loki.delve.office.com/api/v2/person"
+DELV_ORG_URL = "https://eur.loki.delve.office.com/api/v1/organization"
+DELV_WORKINGWITH_URL = "https://eur.loki.delve.office.com/api/v1/workingwith"
+
+
+async def _collect_json_from_responses(page: Page) -> Dict[str, Any]:
+    collected: Dict[str, Any] = {}
+
+    async def handle_response(response: Response):
+        url = response.url
+        try:
+            if url.startswith(DELV_PERSON_URL):
+                collected["person"] = await response.json()
+            elif url.startswith(DELV_ORG_URL):
+                collected["organization"] = await response.json()
+            elif url.startswith(DELV_WORKINGWITH_URL):
+                collected["workingwith"] = await response.json()
+        except Exception:
+            # JSON-Parsing-Fehler ignorieren
+            pass
+
+    page.on("response", handle_response)
+    return collected
+
+
+async def _open_profile_and_collect(page: Page, email: str) -> Dict[str, Any]:
+    collected = await _collect_json_from_responses(page)
+
+    await page.goto("https://m365.cloud.microsoft/search/?auth=2")
+
+    # Suche öffnen
+    search_button = page.get_by_role("button", name="search")
+    await search_button.click()
+
+    input_box = page.locator('input[type="text"]')
+    await input_box.fill(f'Person:"{email}"')
+    await input_box.press("Enter")
+
+    # Warten bis ein Profil-Button erscheint
+    await page.wait_for_timeout(2000)
+    profile_button = page.get_by_title("Organisation")
+    'await profile_button.click()'
+
+    # Organisation-Tab
+    org_button = page.locator('button[data-content="Organisation"]')
+    await org_button.click()
+
+    # Zeit geben, Netzwerk-Calls zu sammeln
+    await page.wait_for_timeout(4000)
+    return collected
+
+
+async def extract_relations_for_manager(manager_email: str) -> List[WorkingWithRelation]:
+    if not AUTH_STATE_FILE.exists():
+        raise RuntimeError("auth_state.json nicht vorhanden, bitte zuerst login-Modus ausführen.")
+
+    async with async_playwright() as p:
+        browser_type = p.chromium
+        context = await browser_type.launch_persistent_context(
+            user_data_dir="user_data",
+            headless=False,
+            channel="chrome"
+        )
+        page = await context.new_page()
+
+        collected = await _open_profile_and_collect(page, manager_email)
+
+        person_json = collected.get("person") or {}
+        org_json = collected.get("organization") or {}
+        working_json = collected.get("workingwith") or {}
+
+        manager_person = parse_person_from_organization_person(person_json)
+        print(manager_person)
+        working_with_persons = parse_workingwith_entries(working_json)
+        print(working_with_persons)
+        directs_emails = parse_direct_emails_from_organization(org_json)
+        print(directs_emails)
+
+        relations: List[WorkingWithRelation] = []
+
+        # Manager -> WorkingWith
+        for dest in working_with_persons:
+            relations.append(WorkingWithRelation(source=manager_person, destination=dest))
+
+        # Für alle Directs ebenfalls WorkingWith holen
+        for direct_email in directs_emails:
+            collected_direct = await _open_profile_and_collect(page, direct_email)
+            person_json_d = collected_direct.get("person") or {}
+            working_json_d = collected_direct.get("workingwith") or {}
+
+            direct_person = parse_person_from_organization_person(person_json_d)
+            working_with_persons_d = parse_workingwith_entries(working_json_d)
+
+            for dest in working_with_persons_d:
+                relations.append(WorkingWithRelation(source=direct_person, destination=dest))
+
+        await context.close()
+        return relations
+
+
+async def extract_relations_for_emails(emails: List[str]) -> List[WorkingWithRelation]:
+    if not AUTH_STATE_FILE.exists():
+        raise RuntimeError("auth_state.json nicht vorhanden – bitte zuerst login-Modus ausführen.")
+
+    async with async_playwright() as p:
+        browser_type = p.chromium
+        context = await browser_type.launch_persistent_context(
+            user_data_dir="user_data",
+            headless=False,
+            channel="chrome"
+        )
+        page = await context.new_page()
+
+        relations: List[WorkingWithRelation] = []
+
+        for email in emails:
+            collected = await _open_profile_and_collect(page, email)
+            person_json = collected.get("person") or {}
+            working_json = collected.get("workingwith") or {}
+
+            person = parse_person_from_organization_person(person_json)
+            working_with_persons = parse_workingwith_entries(working_json)
+
+            for dest in working_with_persons:
+                relations.append(WorkingWithRelation(source=person, destination=dest))
+
+        await context.close()
+        return relations
+
+
+def write_relations_to_csv(relations: List[WorkingWithRelation], output_path: Path) -> None:
+    fieldnames = [
+        "source_mail",
+        "source_displayname",
+        "source_jobTitle",
+        "source_department",
+        "destination_mail",
+        "destination_displayname",
+        "destination_jobTitle",
+        "destination_department",
+    ]
+    with output_path.open("w", newline="", encoding="utf-8") as f:
+        writer = csv.DictWriter(f, fieldnames=fieldnames)
+        writer.writeheader()
+        for rel in relations:
+            writer.writerow(
+                {
+                    "source_mail": rel.source.email,
+                    "source_displayname": rel.source.display_name,
+                    "source_jobTitle": rel.source.job_title,
+                    "source_department": rel.source.department,
+                    "destination_mail": rel.destination.email,
+                    "destination_displayname": rel.destination.display_name,
+                    "destination_jobTitle": rel.destination.job_title,
+                    "destination_department": rel.destination.department,
+                }
+            )
diff --git a/app/main.py b/app/main.py
new file mode 100644
index 0000000..6e6233c
--- /dev/null
+++ b/app/main.py
@@ -0,0 +1,60 @@
+from __future__ import annotations
+import asyncio
+import sys
+from pathlib import Path
+
+from .auth import ensure_login_sync
+from .extractor import (
+    extract_relations_for_manager,
+    extract_relations_for_emails,
+    write_relations_to_csv,
+)
+from .aggregator import aggregate_department_links
+
+
+OUTPUT_DIR = Path("output")
+OUTPUT_DIR.mkdir(exist_ok=True)
+
+
+def main(argv: list[str]) -> None:
+    if len(argv) < 2:
+        print("Verwendung:")
+        print("  python -m app.main login <email>")
+        print("  python -m app.main manager <manager_email>")
+        print("  python -m app.main emails <email1> [<email2> ...]")
+        sys.exit(1)
+
+    mode = argv[1].lower()
+
+    if mode == "login":
+        if len(argv) != 3:
+            print("login-Modus benötigt genau 1 E-Mail-Adresse.")
+            sys.exit(1)
+        email = argv[2]
+        ensure_login_sync(email)
+        return
+
+    elif mode == "manager":
+        if len(argv) != 3:
+            print("manager-Modus benötigt genau 1 Manager-E-Mail.")
+            sys.exit(1)
+        manager_email = argv[2]
+        relations = asyncio.run(extract_relations_for_manager(manager_email))
+
+    elif mode == "emails":
+        if len(argv) < 3:
+            print("emails-Modus benötigt mindestens 1 E-Mail-Adresse.")
+            sys.exit(1)
+        emails = argv[2:]
+        relations = asyncio.run(extract_relations_for_emails(emails))
+
+    else:
+        print(f"Unbekannter Modus: {mode}")
+        sys.exit(1)
+
+    csv_path = OUTPUT_DIR / "relations.csv"
+    write_relations_to_csv(relations, csv_path)
+    print(f"CSV geschrieben: {csv_path}")
+
+if __name__ == "__main__":
+    main(sys.argv)
diff --git a/app/models.py b/app/models.py
new file mode 100644
index 0000000..b281dda
--- /dev/null
+++ b/app/models.py
@@ -0,0 +1,16 @@
+from dataclasses import dataclass
+from typing import Optional
+
+
+@dataclass
+class Person:
+    email: str
+    display_name: Optional[str]
+    job_title: Optional[str]
+    department: Optional[str]
+
+
+@dataclass
+class WorkingWithRelation:
+    source: Person
+    destination: Person
diff --git a/app/parser.py b/app/parser.py
new file mode 100644
index 0000000..2c7bcdd
--- /dev/null
+++ b/app/parser.py
@@ -0,0 +1,131 @@
+from __future__ import annotations
+from typing import Any, Dict, Optional
+from .models import Person
+
+def _first_or_none(items: Optional[list]) -> Optional[dict]:
+    if not items:
+        return None
+    return items[0] or None
+
+def parse_person_from_organization_person(json_obj: Dict[str, Any]) -> Person:
+    """
+    Erwartet ein Objekt der Form:
+
+    {
+      "person": {
+        "names": [
+          {
+            "value": {
+              "displayName": "Display, Name-MGB",
+              "givenName": "Name",
+              "surname": "Display"
+            },
+            "source": "Organisation"
+          }
+        ],
+        "emailAddresses": [
+          {
+            "value": {
+              "name": "display.name@mgb.ch",
+              "address": "display.name@mgb.ch"
+            },
+            "source": "Organisation"
+          }
+        ],
+        "workDetails": [
+          {
+            "value": {
+              "companyName": "Migros-Genossenschafts-Bund",
+              "jobTitle": "Title",
+              "department": "Dept",
+              "office": "HH-xx"
+            },
+            "source": "Organisation"
+          }
+        ]
+      }
+    }
+    """
+
+    person_data = json_obj.get("person", {})
+
+    # Namen
+    names_entry = _first_or_none(person_data.get("names"))
+    names_value = (names_entry or {}).get("value", {}) if names_entry else {}
+    display_name = names_value.get("displayName")
+
+    # E-Mail
+    mail_entry = _first_or_none(person_data.get("emailAddresses"))
+    mail_value = (mail_entry or {}).get("value", {}) if mail_entry else {}
+    email = mail_value.get("address")
+
+    # WorkDetails: Firma, Job, Abteilung, Büro
+    work_entry = _first_or_none(person_data.get("workDetails"))
+    work_value = (work_entry or {}).get("value", {}) if work_entry else {}
+    company_name = work_value.get("companyName")
+    job_title = work_value.get("jobTitle")
+    department = work_value.get("department")
+    office = work_value.get("office")
+
+    return Person(
+        email=email,
+        display_name=display_name,
+        job_title=job_title,
+        department=department,
+    )
+
+
+def parse_direct_emails_from_organization(json_obj: dict[str, Any]) -> List[str]:
+    """
+    Erwartete Struktur (aus organization_anonymized):
+
+    {
+      "managers": [...],
+      "directs": [
+        {
+          "smtp": "direct.report@mgb.ch",
+          "userPrincipalName": "direct.report@mgb.ch",
+          ...
+        }
+      ]
+    }
+    """
+    directs = json_obj.get("directs", [])
+    emails: List[str] = []
+    for item in directs:
+        email = item.get("smtp") or item.get("userPrincipalName")
+        if email:
+            emails.append(email)
+    return emails
+
+
+def parse_workingwith_entries(json_obj: dict[str, Any]) -> List[Person]:
+    """
+    Erwartete Struktur (aus working_with_anonymized):
+
+    {
+      "value": [
+        {
+          "email": "user@migros.ch",
+          "userPrincipalName": "user@migros.ch",
+          "fullName": "Nachname, Vorname-MIGROS",
+          "jobTitle": "Leitung",
+          "department": "Dept 1",
+          ...
+        }
+      ]
+    }
+    """
+    value = json_obj.get("value", [])
+    persons: List[Person] = []
+    for item in value:
+        email = item.get("email") or item.get("userPrincipalName")
+        persons.append(
+            Person(
+                email=email,
+                display_name=item.get("fullName"),
+                job_title=item.get("jobTitle"),
+                department=item.get("department"),
+            )
+        )
+    return persons
diff --git a/install.bat b/install.bat
new file mode 100644
index 0000000..cf3bfc9
--- /dev/null
+++ b/install.bat
@@ -0,0 +1,18 @@
+@echo off
+setlocal
+
+REM Virtuelle Umgebung erstellen
+python -m venv venv
+
+REM venv aktivieren
+call venv\Scripts\activate.bat
+
+REM Abhängigkeiten installieren
+pip install --upgrade pip
+pip install -r requirements.txt
+
+REM Playwright-Browser installieren
+playwright install chromium
+
+echo Installation abgeschlossen.
+endlocal
diff --git a/requirements.txt b/requirements.txt
new file mode 100644
index 0000000..1a6eed5
--- /dev/null
+++ b/requirements.txt
@@ -0,0 +1,4 @@
+playwright
+pyvis
+plotly
+pandas
\ No newline at end of file
diff --git a/run.bat b/run.bat
new file mode 100644
index 0000000..260cd8b
--- /dev/null
+++ b/run.bat
@@ -0,0 +1,17 @@
+@echo off
+setlocal
+
+REM Virtuelle Umgebung aktivieren
+call venv\Scripts\activate.bat
+
+if "%1"=="" (
+    echo Nutzung:
+    echo   run.bat login ^<email^>
+    echo   run.bat manager ^<manager_email^>
+    echo   run.bat emails ^<email1^> [^<email2^> ...]
+    goto :eof
+)
+
+python -m app.main %*
+
+endlocal
diff --git a/visualize.bat b/visualize.bat
new file mode 100644
index 0000000..04a095e
--- /dev/null
+++ b/visualize.bat
@@ -0,0 +1,15 @@
+@echo off
+setlocal
+
+REM Virtuelle Umgebung aktivieren
+call venv\Scripts\activate.bat
+
+if "%1"=="" (
+    echo Nutzung:
+    echo   visualize.bat ^<path_to_csv^>
+    goto :eof
+)
+
+python -m app.easyvisualize.py %*
+
+endlocal