Upload 26 files

Dockerfile

@@ -0,0 +1,22 @@
+FROM python:3.11-slim
+
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
+ENV PORT=7860
+ENV PYTHONPATH=/app
+
+WORKDIR /app
+
+RUN apt-get update && apt-get install -y --no-install-recommends     build-essential     curl     && rm -rf /var/lib/apt/lists/*
+
+COPY requirements.txt .
+RUN pip install --no-cache-dir -r requirements.txt
+
+COPY backend /app/backend
+COPY data_samples /app/data_samples
+
+RUN mkdir -p /data/uploads /data/artifacts
+
+EXPOSE 7860
+
+CMD ["uvicorn", "backend.app.main:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -0,0 +1,34 @@
+---
+title: ClusterBuster API
+emoji: 📊
+colorFrom: blue
+colorTo: indigo
+sdk: docker
+app_port: 7860
+pinned: false
+---
+
+# ClusterBuster Multi Algorithm
+
+Backend:
+- FastAPI
+- SQLite at /data/clusterforge.db
+- Upload datasets
+- Profile datasets
+- Run clustering experiments
+- PCA scatter points
+- Runs and compare APIs
+- CSV export
+- Multi algorithm clustering
+
+Frontend:
+- Next.js
+- Upload page
+- New experiment page
+- Runs page
+- Compare page
+- PCA scatter plot
+
+Deploy:
+- backend repo to Hugging Face Docker Space
+- frontend repo to Vercel

backend/app/api/compare.py

@@ -0,0 +1,35 @@
+import json
+
+from fastapi import APIRouter, Depends, Query
+from sqlalchemy.orm import Session
+
+from backend.app.db import get_db
+from backend.app.repositories.experiment_repo import get_experiments_by_ids, list_experiments
+
+router = APIRouter(tags=["compare"])
+
+
+@router.get("/compare")
+def compare_experiments(experiment_ids: str | None = Query(default=None), db: Session = Depends(get_db)):
+    if experiment_ids:
+        ids = [x.strip() for x in experiment_ids.split(",") if x.strip()]
+        experiments = get_experiments_by_ids(db, ids)
+    else:
+        experiments = list_experiments(db)
+
+    return {
+        "experiments": [
+            {
+                "experiment_id": exp.id,
+                "dataset_id": exp.dataset_id,
+                "algorithm": exp.algorithm,
+                "status": exp.status,
+                "config": json.loads(exp.config_json) if exp.config_json else {},
+                "metrics": json.loads(exp.metrics_json) if exp.metrics_json else {},
+                "summary": json.loads(exp.summary_json) if exp.summary_json else {},
+                "runtime_ms": exp.runtime_ms,
+                "error_message": exp.error_message,
+            }
+            for exp in experiments
+        ]
+    }

backend/app/api/datasets.py

@@ -0,0 +1,89 @@
+import json
+from pathlib import Path
+
+import pandas as pd
+from fastapi import APIRouter, Depends, File, HTTPException, UploadFile
+from sqlalchemy.orm import Session
+
+from backend.app.db import get_db
+from backend.app.repositories.dataset_repo import create_dataset, get_dataset, list_datasets
+from backend.app.services.profiling_service import profile_dataframe
+from backend.app.utils.ids import make_dataset_id
+
+router = APIRouter(tags=["datasets"])
+
+UPLOAD_DIR = Path("/data/uploads")
+UPLOAD_DIR.mkdir(parents=True, exist_ok=True)
+
+
+@router.get("/datasets")
+def datasets_list(db: Session = Depends(get_db)):
+    datasets = list_datasets(db)
+    return {
+        "datasets": [
+            {
+                "id": d.id,
+                "name": d.name,
+                "row_count": d.row_count,
+                "column_count": d.column_count,
+            }
+            for d in datasets
+        ]
+    }
+
+
+@router.post("/datasets/upload")
+async def upload_dataset(file: UploadFile = File(...), db: Session = Depends(get_db)):
+    if not file.filename:
+        raise HTTPException(status_code=400, detail="Missing file name")
+
+    suffix = Path(file.filename).suffix.lower()
+    if suffix not in {".csv", ".xlsx", ".xls"}:
+        raise HTTPException(status_code=400, detail="Only CSV and Excel files are supported")
+
+    dataset_id = make_dataset_id()
+    path = UPLOAD_DIR / f"{dataset_id}{suffix}"
+
+    content = await file.read()
+    path.write_bytes(content)
+
+    if suffix == ".csv":
+        df = pd.read_csv(path)
+    else:
+        df = pd.read_excel(path)
+
+    profile = profile_dataframe(df)
+
+    create_dataset(
+        db=db,
+        id=dataset_id,
+        name=file.filename,
+        file_path=str(path),
+        row_count=int(len(df)),
+        column_count=int(len(df.columns)),
+        schema_json=json.dumps({"columns": list(df.columns)}),
+        profile_json=json.dumps(profile),
+    )
+
+    return {
+        "dataset_id": dataset_id,
+        "name": file.filename,
+        "row_count": int(len(df)),
+        "column_count": int(len(df.columns)),
+    }
+
+
+@router.get("/datasets/{dataset_id}/profile")
+def dataset_profile(dataset_id: str, db: Session = Depends(get_db)):
+    dataset = get_dataset(db, dataset_id)
+    if not dataset:
+        raise HTTPException(status_code=404, detail="Dataset not found")
+
+    return {
+        "dataset_id": dataset.id,
+        "name": dataset.name,
+        "row_count": dataset.row_count,
+        "column_count": dataset.column_count,
+        "schema": json.loads(dataset.schema_json),
+        "profile": json.loads(dataset.profile_json),
+    }

backend/app/api/experiments.py

@@ -0,0 +1,240 @@
+import json
+import time
+
+import pandas as pd
+from fastapi import APIRouter, Depends, HTTPException
+from pydantic import BaseModel
+from sqlalchemy.orm import Session
+from sklearn.cluster import (
+    AffinityPropagation,
+    AgglomerativeClustering,
+    Birch,
+    BisectingKMeans,
+    DBSCAN,
+    KMeans,
+    MeanShift,
+    MiniBatchKMeans,
+    OPTICS,
+    SpectralClustering,
+)
+from sklearn.decomposition import PCA
+from sklearn.metrics import silhouette_score
+from sklearn.mixture import GaussianMixture
+
+from backend.app.db import get_db
+from backend.app.repositories.dataset_repo import get_dataset
+from backend.app.repositories.experiment_repo import create_experiment
+from backend.app.utils.ids import make_experiment_id
+
+try:
+    import hdbscan
+except ImportError:
+    hdbscan = None
+
+router = APIRouter(tags=["experiments"])
+
+
+class RunRequest(BaseModel):
+    dataset_id: str
+    name: str | None = None
+    algorithm: str = "kmeans"
+    n_clusters: int = 4
+    feature_columns: list[str]
+    algorithm_params: dict | None = None
+
+
+def build_model(algorithm: str, n_clusters: int, params: dict):
+    if algorithm == "kmeans":
+        return KMeans(
+            n_clusters=params.get("n_clusters", n_clusters),
+            n_init=10,
+            random_state=42,
+        ), "fit_predict"
+
+    if algorithm == "mini_batch_kmeans":
+        return MiniBatchKMeans(
+            n_clusters=params.get("n_clusters", n_clusters),
+            random_state=42,
+            batch_size=params.get("batch_size", 1024),
+        ), "fit_predict"
+
+    if algorithm == "agglomerative":
+        return AgglomerativeClustering(
+            n_clusters=params.get("n_clusters", n_clusters),
+            linkage=params.get("linkage", "ward"),
+        ), "fit_predict"
+
+    if algorithm == "birch":
+        return Birch(
+            n_clusters=params.get("n_clusters", n_clusters),
+            threshold=params.get("threshold", 0.5),
+        ), "fit_predict"
+
+    if algorithm == "dbscan":
+        return DBSCAN(
+            eps=params.get("eps", 0.5),
+            min_samples=params.get("min_samples", 5),
+        ), "fit_predict"
+
+    if algorithm == "optics":
+        return OPTICS(
+            min_samples=params.get("min_samples", 5),
+        ), "fit_predict"
+
+    if algorithm == "mean_shift":
+        return MeanShift(), "fit_predict"
+
+    if algorithm == "spectral":
+        return SpectralClustering(
+            n_clusters=params.get("n_clusters", n_clusters),
+            random_state=42,
+            assign_labels="kmeans",
+            affinity=params.get("affinity", "nearest_neighbors"),
+        ), "fit_predict"
+
+    if algorithm == "gaussian_mixture":
+        return GaussianMixture(
+            n_components=params.get("n_components", n_clusters),
+            random_state=42,
+        ), "predict_after_fit"
+
+    if algorithm == "affinity_propagation":
+        return AffinityPropagation(random_state=42), "fit_predict"
+
+    if algorithm == "bisecting_kmeans":
+        return BisectingKMeans(
+            n_clusters=params.get("n_clusters", n_clusters),
+            random_state=42,
+        ), "fit_predict"
+
+    if algorithm == "hdbscan":
+        if hdbscan is None:
+            raise HTTPException(status_code=400, detail="HDBSCAN is not installed")
+        return hdbscan.HDBSCAN(
+            min_cluster_size=params.get("min_cluster_size", 10),
+            min_samples=params.get("min_samples", 5),
+        ), "fit_predict"
+
+    raise HTTPException(status_code=400, detail="Unsupported algorithm")
+
+
+@router.post("/experiments/run")
+def run_experiment(req: RunRequest, db: Session = Depends(get_db)):
+    dataset = get_dataset(db, req.dataset_id)
+    if not dataset:
+        raise HTTPException(status_code=404, detail="Dataset not found")
+
+    if dataset.file_path.endswith(".csv"):
+        df = pd.read_csv(dataset.file_path)
+    else:
+        df = pd.read_excel(dataset.file_path)
+
+    if not req.feature_columns:
+        raise HTTPException(status_code=400, detail="feature_columns is required")
+
+    missing = [c for c in req.feature_columns if c not in df.columns]
+    if missing:
+        raise HTTPException(status_code=400, detail=f"Missing columns: {', '.join(missing)}")
+
+    X = df[req.feature_columns].copy()
+    X = pd.get_dummies(X)
+    X = X.fillna(0)
+
+    params = req.algorithm_params or {}
+    start = time.time()
+
+    model, mode = build_model(req.algorithm, req.n_clusters, params)
+
+    if mode == "fit_predict":
+        labels = model.fit_predict(X)
+    elif mode == "predict_after_fit":
+        model.fit(X)
+        labels = model.predict(X)
+    else:
+        raise HTTPException(status_code=400, detail="Invalid model execution mode")
+
+    runtime_ms = int((time.time() - start) * 1000)
+    unique_labels = sorted(set(labels.tolist()))
+
+    score = None
+    valid_labels = [x for x in unique_labels if x != -1]
+    if len(valid_labels) > 1 and len(valid_labels) < len(X):
+        try:
+            score = float(silhouette_score(X, labels))
+        except Exception:
+            score = None
+
+    pca = PCA(n_components=2, random_state=42)
+    coords = pca.fit_transform(X)
+
+    points = [
+        {
+            "row_index": int(i),
+            "cluster_label": int(labels[i]),
+            "x": float(coords[i][0]),
+            "y": float(coords[i][1]),
+        }
+        for i in range(len(labels))
+    ]
+
+    cluster_sizes = {str(label): int((labels == label).sum()) for label in unique_labels}
+    experiment_id = make_experiment_id()
+
+    metrics = {
+        "silhouette_score": score,
+        "cluster_count": len(valid_labels) if valid_labels else len(unique_labels),
+        "row_count": int(len(X)),
+        "runtime_ms": runtime_ms,
+        "noise_count": int((labels == -1).sum()) if -1 in unique_labels else 0,
+    }
+
+    summary = {
+        "feature_columns": req.feature_columns,
+        "cluster_sizes": cluster_sizes,
+        "points": points,
+    }
+
+    create_experiment(
+        db=db,
+        id=experiment_id,
+        dataset_id=req.dataset_id,
+        algorithm=req.algorithm,
+        status="completed",
+        config_json=json.dumps(req.model_dump()),
+        metrics_json=json.dumps(metrics),
+        summary_json=json.dumps(summary),
+        runtime_ms=runtime_ms,
+        error_message=None,
+    )
+
+    return {
+        "experiment_id": experiment_id,
+        "status": "completed",
+        "silhouette_score": score,
+        "cluster_count": metrics["cluster_count"],
+        "cluster_sizes": cluster_sizes,
+        "runtime_ms": runtime_ms,
+        "noise_count": metrics["noise_count"],
+        "points": points,
+    }
+
+
+@router.get("/experiments/{experiment_id}/results")
+def experiment_results(experiment_id: str, db: Session = Depends(get_db)):
+    from backend.app.repositories.experiment_repo import get_experiment
+
+    exp = get_experiment(db, experiment_id)
+    if not exp:
+        raise HTTPException(status_code=404, detail="Experiment not found")
+
+    return {
+        "experiment_id": exp.id,
+        "dataset_id": exp.dataset_id,
+        "algorithm": exp.algorithm,
+        "status": exp.status,
+        "config": json.loads(exp.config_json) if exp.config_json else {},
+        "metrics": json.loads(exp.metrics_json) if exp.metrics_json else {},
+        "summary": json.loads(exp.summary_json) if exp.summary_json else {},
+        "runtime_ms": exp.runtime_ms,
+        "error_message": exp.error_message,
+    }

backend/app/api/exports.py

@@ -0,0 +1,51 @@
+import io
+import json
+
+import pandas as pd
+from fastapi import APIRouter, Depends, HTTPException
+from fastapi.responses import StreamingResponse
+from sqlalchemy.orm import Session
+
+from backend.app.db import get_db
+from backend.app.repositories.dataset_repo import get_dataset
+from backend.app.repositories.experiment_repo import get_experiment
+
+router = APIRouter(tags=["exports"])
+
+
+@router.get("/exports/{experiment_id}")
+def export_experiment(experiment_id: str, db: Session = Depends(get_db)):
+    experiment = get_experiment(db, experiment_id)
+    if not experiment:
+        raise HTTPException(status_code=404, detail="Experiment not found")
+
+    dataset = get_dataset(db, experiment.dataset_id)
+    if not dataset:
+        raise HTTPException(status_code=404, detail="Dataset not found")
+
+    if dataset.file_path.endswith(".csv"):
+        df = pd.read_csv(dataset.file_path)
+    else:
+        df = pd.read_excel(dataset.file_path)
+
+    summary = json.loads(experiment.summary_json) if experiment.summary_json else {}
+    points = summary.get("points", [])
+    export_df = df.copy()
+    if points and len(points) == len(df):
+        export_df["cluster_label"] = [p["cluster_label"] for p in points]
+        export_df["pca_x"] = [p["x"] for p in points]
+        export_df["pca_y"] = [p["y"] for p in points]
+
+    metrics = json.loads(experiment.metrics_json) if experiment.metrics_json else {}
+    for key, value in metrics.items():
+        export_df[f"metric_{key}"] = value
+
+    buffer = io.StringIO()
+    export_df.to_csv(buffer, index=False)
+    buffer.seek(0)
+
+    return StreamingResponse(
+        iter([buffer.getvalue()]),
+        media_type="text/csv",
+        headers={"Content-Disposition": f"attachment; filename={experiment_id}_export.csv"},
+    )

backend/app/api/health.py

@@ -0,0 +1,7 @@
+from fastapi import APIRouter
+
+router = APIRouter(tags=["health"])
+
+@router.get("/health")
+def health():
+    return {"ok": True, "service": "clusterbuster-api"}

backend/app/api/presets.py

@@ -0,0 +1,23 @@
+from fastapi import APIRouter
+
+router = APIRouter(tags=["presets"])
+
+
+@router.get("/presets/algorithms")
+def list_algorithms():
+    return {
+        "algorithms": [
+            {"key": "kmeans", "label": "KMeans", "params": {"n_clusters": 4}},
+            {"key": "mini_batch_kmeans", "label": "MiniBatchKMeans", "params": {"n_clusters": 4, "batch_size": 1024}},
+            {"key": "agglomerative", "label": "Agglomerative", "params": {"n_clusters": 4}},
+            {"key": "birch", "label": "Birch", "params": {"n_clusters": 4, "threshold": 0.5}},
+            {"key": "dbscan", "label": "DBSCAN", "params": {"eps": 0.5, "min_samples": 5}},
+            {"key": "optics", "label": "OPTICS", "params": {"min_samples": 5}},
+            {"key": "mean_shift", "label": "Mean Shift", "params": {}},
+            {"key": "spectral", "label": "Spectral Clustering", "params": {"n_clusters": 4, "affinity": "nearest_neighbors"}},
+            {"key": "gaussian_mixture", "label": "Gaussian Mixture", "params": {"n_components": 4}},
+            {"key": "affinity_propagation", "label": "Affinity Propagation", "params": {}},
+            {"key": "bisecting_kmeans", "label": "Bisecting KMeans", "params": {"n_clusters": 4}},
+            {"key": "hdbscan", "label": "HDBSCAN", "params": {"min_cluster_size": 10, "min_samples": 5}}
+        ]
+    }

backend/app/api/runs.py

@@ -0,0 +1,29 @@
+import json
+
+from fastapi import APIRouter, Depends
+from sqlalchemy.orm import Session
+
+from backend.app.db import get_db
+from backend.app.repositories.experiment_repo import list_experiments
+
+router = APIRouter(tags=["runs"])
+
+
+@router.get("/runs")
+def get_runs(db: Session = Depends(get_db)):
+    experiments = list_experiments(db)
+    return {
+        "runs": [
+            {
+                "experiment_id": exp.id,
+                "dataset_id": exp.dataset_id,
+                "algorithm": exp.algorithm,
+                "status": exp.status,
+                "metrics": json.loads(exp.metrics_json) if exp.metrics_json else {},
+                "summary": json.loads(exp.summary_json) if exp.summary_json else {},
+                "runtime_ms": exp.runtime_ms,
+                "error_message": exp.error_message,
+            }
+            for exp in experiments
+        ]
+    }

backend/app/db.py

@@ -0,0 +1,34 @@
+import os
+from sqlalchemy import create_engine
+from sqlalchemy.orm import declarative_base, sessionmaker
+
+DB_PATH = os.getenv("DB_PATH", "/data/clusterforge.db")
+DATABASE_URL = f"sqlite:///{DB_PATH}"
+
+engine = create_engine(
+    DATABASE_URL,
+    connect_args={"check_same_thread": False},
+)
+
+SessionLocal = sessionmaker(
+    autocommit=False,
+    autoflush=False,
+    bind=engine,
+)
+
+Base = declarative_base()
+
+
+def init_db() -> None:
+    import backend.app.models.dataset
+    import backend.app.models.experiment
+
+    Base.metadata.create_all(bind=engine)
+
+
+def get_db():
+    db = SessionLocal()
+    try:
+        yield db
+    finally:
+        db.close()

backend/app/main.py

@@ -0,0 +1,45 @@
+from fastapi import FastAPI
+from fastapi.middleware.cors import CORSMiddleware
+
+from backend.app.db import init_db
+from backend.app.api.health import router as health_router
+from backend.app.api.datasets import router as datasets_router
+from backend.app.api.presets import router as presets_router
+from backend.app.api.experiments import router as experiments_router
+from backend.app.api.compare import router as compare_router
+from backend.app.api.exports import router as exports_router
+from backend.app.api.runs import router as runs_router
+
+app = FastAPI(title="ClusterBuster API")
+
+
+@app.on_event("startup")
+def on_startup() -> None:
+    init_db()
+
+
+origins = [
+    "http://localhost:3000",
+    "https://cluster-buster.vercel.app",
+]
+
+app.add_middleware(
+    CORSMiddleware,
+    allow_origins=origins,
+    allow_credentials=True,
+    allow_methods=["*"],
+    allow_headers=["*"],
+)
+
+app.include_router(health_router, prefix="/api")
+app.include_router(datasets_router, prefix="/api")
+app.include_router(presets_router, prefix="/api")
+app.include_router(experiments_router, prefix="/api")
+app.include_router(compare_router, prefix="/api")
+app.include_router(exports_router, prefix="/api")
+app.include_router(runs_router, prefix="/api")
+
+
+@app.get("/")
+def root():
+    return {"ok": True, "service": "clusterbuster-api"}

backend/app/models/dataset.py

@@ -0,0 +1,14 @@
+from sqlalchemy import Column, Integer, String, Text
+from backend.app.db import Base
+
+
+class Dataset(Base):
+    __tablename__ = "datasets"
+
+    id = Column(String, primary_key=True, index=True)
+    name = Column(String, nullable=False)
+    file_path = Column(String, nullable=False)
+    row_count = Column(Integer, nullable=False)
+    column_count = Column(Integer, nullable=False)
+    schema_json = Column(Text, nullable=False)
+    profile_json = Column(Text, nullable=False)

backend/app/models/experiment.py

@@ -0,0 +1,16 @@
+from sqlalchemy import Column, Integer, String, Text
+from backend.app.db import Base
+
+
+class Experiment(Base):
+    __tablename__ = "experiments"
+
+    id = Column(String, primary_key=True, index=True)
+    dataset_id = Column(String, nullable=False, index=True)
+    algorithm = Column(String, nullable=False)
+    status = Column(String, nullable=False)
+    config_json = Column(Text, nullable=True)
+    metrics_json = Column(Text, nullable=True)
+    summary_json = Column(Text, nullable=True)
+    runtime_ms = Column(Integer, nullable=True)
+    error_message = Column(Text, nullable=True)

backend/app/repositories/dataset_repo.py

@@ -0,0 +1,35 @@
+from sqlalchemy.orm import Session
+from backend.app.models.dataset import Dataset
+
+
+def create_dataset(
+    db: Session,
+    id: str,
+    name: str,
+    file_path: str,
+    row_count: int,
+    column_count: int,
+    schema_json: str,
+    profile_json: str,
+) -> Dataset:
+    dataset = Dataset(
+        id=id,
+        name=name,
+        file_path=file_path,
+        row_count=row_count,
+        column_count=column_count,
+        schema_json=schema_json,
+        profile_json=profile_json,
+    )
+    db.add(dataset)
+    db.commit()
+    db.refresh(dataset)
+    return dataset
+
+
+def get_dataset(db: Session, dataset_id: str) -> Dataset | None:
+    return db.query(Dataset).filter(Dataset.id == dataset_id).first()
+
+
+def list_datasets(db: Session) -> list[Dataset]:
+    return db.query(Dataset).order_by(Dataset.name.asc()).all()

backend/app/repositories/experiment_repo.py

@@ -0,0 +1,45 @@
+from sqlalchemy.orm import Session
+from backend.app.models.experiment import Experiment
+
+
+def create_experiment(
+    db: Session,
+    id: str,
+    dataset_id: str,
+    algorithm: str,
+    status: str,
+    config_json: str | None = None,
+    metrics_json: str | None = None,
+    summary_json: str | None = None,
+    runtime_ms: int | None = None,
+    error_message: str | None = None,
+) -> Experiment:
+    experiment = Experiment(
+        id=id,
+        dataset_id=dataset_id,
+        algorithm=algorithm,
+        status=status,
+        config_json=config_json,
+        metrics_json=metrics_json,
+        summary_json=summary_json,
+        runtime_ms=runtime_ms,
+        error_message=error_message,
+    )
+    db.add(experiment)
+    db.commit()
+    db.refresh(experiment)
+    return experiment
+
+
+def get_experiment(db: Session, experiment_id: str) -> Experiment | None:
+    return db.query(Experiment).filter(Experiment.id == experiment_id).first()
+
+
+def list_experiments(db: Session) -> list[Experiment]:
+    return db.query(Experiment).order_by(Experiment.id.desc()).all()
+
+
+def get_experiments_by_ids(db: Session, experiment_ids: list[str]) -> list[Experiment]:
+    if not experiment_ids:
+        return []
+    return db.query(Experiment).filter(Experiment.id.in_(experiment_ids)).all()

backend/app/services/profiling_service.py

@@ -0,0 +1,23 @@
+def profile_dataframe(df):
+    numeric_cols = df.select_dtypes(include=["int64", "float64", "int32", "float32"]).columns.tolist()
+    categorical_cols = df.select_dtypes(include=["object", "bool"]).columns.tolist()
+
+    recommended = []
+    if numeric_cols:
+        recommended.extend(["kmeans", "mini_batch_kmeans", "birch", "bisecting_kmeans", "gaussian_mixture"])
+    if categorical_cols:
+        recommended.extend(["agglomerative", "dbscan", "optics", "spectral", "hdbscan"])
+
+    cols = []
+    for col in df.columns:
+        cols.append({
+            "name": col,
+            "inferred_type": str(df[col].dtype),
+            "missing_pct": float(df[col].isna().mean()),
+            "cardinality": int(df[col].nunique(dropna=True)),
+        })
+
+    return {
+        "columns": cols,
+        "recommended_algorithms": recommended,
+    }

backend/app/utils/ids.py

@@ -0,0 +1,7 @@
+import secrets
+
+def make_dataset_id() -> str:
+    return "ds_" + secrets.token_hex(4)
+
+def make_experiment_id() -> str:
+    return "exp_" + secrets.token_hex(4)

data_samples/access_demo.csv

@@ -0,0 +1,801 @@
+user_id,department,title,location,application_name,entitlement_count,privileged_access_flag,mfa_enabled,risk_score,peer_access_overlap_score,last_login_days_ago,granted_count_6m,revoked_count_6m,review_count_6m,sensitive_access_flag
+user_0,Engineering,Admin,SF,SAP,26,0,1,0.765,0.798,3,19,1,4,0
+user_1,Sales,Specialist,Remote,Oracle,11,1,1,0.377,0.531,37,1,5,8,0
+user_2,Finance,Engineer,NY,Workday,27,0,1,0.807,0.237,40,16,8,6,1
+user_3,HR,Manager,NY,Workday,20,0,1,0.613,0.878,7,17,4,12,0
+user_4,Engineering,Manager,Remote,Jira,21,1,0,0.611,0.231,178,13,0,5,1
+user_5,Sales,Engineer,Remote,GitHub,12,0,1,0.652,0.203,180,7,0,10,1
+user_6,HR,Manager,NY,SAP,15,0,1,0.81,0.448,91,13,0,2,1
+user_7,HR,Engineer,SF,GitHub,10,0,0,0.651,0.457,172,3,0,6,1
+user_8,Sales,Manager,SF,SAP,12,0,1,0.852,0.489,4,0,10,3,0
+user_9,Support,Manager,NY,SAP,43,1,0,0.724,0.607,65,9,10,7,0
+user_10,Engineering,Analyst,Remote,Workday,35,1,0,0.377,0.2,147,20,9,9,1
+user_11,Support,Admin,TX,Jira,5,1,1,0.413,0.318,43,18,0,3,0
+user_12,Sales,Manager,SF,Jira,34,1,0,0.683,0.938,166,7,7,1,1
+user_13,Finance,Manager,TX,Jira,14,1,1,0.666,0.848,150,0,5,8,0
+user_14,Support,Admin,Remote,Oracle,20,1,0,0.264,0.103,178,11,8,8,0
+user_15,Engineering,Admin,TX,Oracle,32,1,0,0.868,0.902,23,14,3,3,0
+user_16,Sales,Specialist,Remote,Workday,41,1,1,0.355,0.453,170,14,9,9,0
+user_17,Finance,Manager,SF,SAP,45,1,0,0.697,0.947,161,8,0,7,0
+user_18,Sales,Analyst,SF,Oracle,18,0,0,0.84,0.851,104,14,4,6,0
+user_19,Support,Engineer,NY,SAP,42,1,1,0.748,0.599,142,7,8,9,0
+user_20,Support,Specialist,TX,Oracle,25,1,0,0.705,0.15,145,10,6,5,1
+user_21,Support,Specialist,Remote,GitHub,11,0,1,0.842,0.609,31,9,9,1,1
+user_22,Sales,Admin,TX,Oracle,27,1,0,0.919,0.189,32,13,1,11,0
+user_23,Sales,Engineer,TX,Oracle,25,0,0,0.924,0.607,55,5,4,1,1
+user_24,Finance,Engineer,NY,SAP,6,0,0,0.715,0.492,163,5,9,2,0
+user_25,Support,Manager,NY,Oracle,13,0,1,0.303,0.549,100,5,2,1,0
+user_26,Engineering,Analyst,Remote,Oracle,6,0,1,0.34,0.279,143,6,3,3,0
+user_27,Engineering,Analyst,SF,Oracle,20,0,0,0.92,0.374,159,3,1,7,0
+user_28,Finance,Specialist,SF,GitHub,21,1,1,0.221,0.965,135,14,0,12,0
+user_29,HR,Analyst,SF,Oracle,29,0,1,0.95,0.693,135,0,2,9,1
+user_30,Sales,Engineer,Remote,Oracle,26,0,1,0.715,0.709,30,20,7,10,0
+user_31,Engineering,Engineer,SF,SAP,9,0,1,0.738,0.235,140,16,4,11,1
+user_32,HR,Engineer,TX,Jira,6,1,1,0.61,0.504,136,16,5,5,0
+user_33,Sales,Admin,SF,Oracle,42,0,0,0.371,0.691,139,0,0,3,0
+user_34,HR,Analyst,NY,Oracle,15,0,0,0.367,0.421,79,3,7,6,0
+user_35,Sales,Specialist,TX,SAP,28,1,0,0.833,0.317,89,8,0,12,0
+user_36,Sales,Analyst,SF,Oracle,23,1,0,0.643,0.479,125,9,2,12,1
+user_37,Sales,Analyst,TX,Jira,9,1,1,0.245,0.476,37,12,7,6,0
+user_38,Sales,Specialist,Remote,Oracle,14,1,1,0.229,0.767,174,10,7,3,1
+user_39,Engineering,Admin,SF,GitHub,35,1,1,0.46,0.101,57,1,6,10,1
+user_40,Support,Admin,NY,SAP,43,1,1,0.946,0.904,97,13,9,4,0
+user_41,Sales,Admin,SF,SAP,36,0,1,0.82,0.923,174,16,7,1,0
+user_42,Sales,Specialist,Remote,SAP,29,1,0,0.305,0.286,11,16,8,1,1
+user_43,Support,Manager,TX,SAP,14,0,0,0.72,0.557,89,18,2,2,0
+user_44,Support,Manager,NY,SAP,34,1,1,0.759,0.27,116,0,5,8,0
+user_45,Engineering,Analyst,Remote,Oracle,5,0,1,0.707,0.578,150,9,7,9,1
+user_46,Engineering,Engineer,Remote,Workday,26,0,1,0.798,0.664,143,3,5,4,0
+user_47,Engineering,Engineer,SF,Jira,30,1,1,0.864,0.135,83,19,6,2,0
+user_48,Finance,Specialist,Remote,Jira,16,0,1,0.553,0.953,3,19,6,0,0
+user_49,HR,Specialist,Remote,GitHub,12,1,1,0.372,0.712,66,14,4,9,0
+user_50,HR,Analyst,TX,Workday,20,1,0,0.817,0.818,91,0,8,0,0
+user_51,HR,Analyst,SF,SAP,36,0,1,0.591,0.776,119,8,4,3,0
+user_52,Finance,Engineer,NY,SAP,30,0,0,0.672,0.822,33,10,0,4,0
+user_53,Support,Manager,TX,Oracle,10,1,1,0.302,0.819,54,17,0,8,1
+user_54,Engineering,Manager,NY,Jira,42,1,1,0.901,0.6,91,19,3,10,1
+user_55,Engineering,Manager,Remote,Workday,41,1,1,0.552,0.34,49,16,6,3,1
+user_56,Support,Admin,TX,Oracle,21,1,1,0.754,0.359,30,7,7,0,1
+user_57,HR,Analyst,TX,GitHub,28,0,0,0.666,0.409,112,12,10,1,1
+user_58,Support,Engineer,TX,SAP,22,0,1,0.889,0.952,40,18,5,8,0
+user_59,Finance,Analyst,NY,SAP,9,1,0,0.912,0.358,8,16,7,10,0
+user_60,Support,Manager,NY,SAP,23,0,0,0.518,0.391,166,18,3,9,1
+user_61,Finance,Admin,SF,GitHub,20,1,0,0.576,0.236,55,4,5,8,1
+user_62,Support,Admin,Remote,Oracle,43,0,0,0.721,0.693,140,15,6,4,0
+user_63,Finance,Specialist,Remote,SAP,44,1,1,0.78,0.904,5,16,10,11,1
+user_64,Engineering,Engineer,SF,SAP,21,1,0,0.934,0.489,65,15,4,4,0
+user_65,Engineering,Engineer,NY,SAP,31,1,1,0.689,0.757,144,6,1,6,1
+user_66,HR,Specialist,SF,GitHub,15,1,1,0.692,0.576,62,14,5,6,0
+user_67,HR,Specialist,TX,Oracle,43,1,0,0.437,0.913,75,17,10,7,1
+user_68,Support,Engineer,Remote,SAP,45,1,1,0.75,0.724,176,19,3,1,1
+user_69,Finance,Engineer,Remote,Workday,20,0,0,0.283,0.729,17,0,10,4,0
+user_70,Finance,Admin,NY,Jira,5,0,0,0.471,0.634,87,6,1,8,0
+user_71,HR,Specialist,Remote,SAP,26,0,0,0.896,0.241,77,1,0,10,1
+user_72,Sales,Manager,TX,Jira,34,1,1,0.92,0.875,49,1,0,5,0
+user_73,Support,Specialist,Remote,Workday,19,1,0,0.935,0.526,74,19,4,8,0
+user_74,HR,Admin,Remote,GitHub,8,1,0,0.91,0.569,31,19,1,12,1
+user_75,Finance,Engineer,Remote,GitHub,44,0,0,0.391,0.741,168,4,7,10,1
+user_76,Finance,Admin,SF,Workday,30,1,1,0.628,0.277,38,2,6,0,1
+user_77,Engineering,Engineer,TX,GitHub,42,0,0,0.595,0.393,21,14,6,0,1
+user_78,Sales,Engineer,NY,Oracle,5,1,0,0.712,0.161,57,12,7,9,1
+user_79,HR,Manager,Remote,Oracle,38,1,1,0.231,0.648,107,11,9,4,1
+user_80,Finance,Admin,NY,GitHub,20,1,1,0.26,0.623,6,15,2,7,1
+user_81,HR,Engineer,Remote,GitHub,30,1,0,0.477,0.756,15,7,5,12,0
+user_82,Sales,Specialist,Remote,Workday,37,1,0,0.547,0.449,27,12,0,3,1
+user_83,Support,Analyst,NY,Oracle,32,0,1,0.569,0.823,138,18,6,11,1
+user_84,HR,Manager,SF,Jira,35,0,0,0.43,0.895,75,15,6,9,0
+user_85,Engineering,Specialist,Remote,Jira,43,0,0,0.856,0.276,34,1,0,1,1
+user_86,Support,Admin,TX,GitHub,11,0,1,0.37,0.133,99,0,8,10,1
+user_87,HR,Admin,SF,SAP,36,1,1,0.908,0.896,71,13,6,5,0
+user_88,Sales,Specialist,NY,Workday,43,0,1,0.234,0.496,106,16,9,5,1
+user_89,HR,Manager,TX,Oracle,40,0,1,0.83,0.527,117,9,5,5,1
+user_90,Support,Specialist,SF,Oracle,40,1,1,0.373,0.869,94,2,0,7,1
+user_91,Finance,Engineer,Remote,Jira,27,1,1,0.383,0.346,76,15,2,2,0
+user_92,HR,Admin,TX,GitHub,24,1,0,0.737,0.89,166,18,7,1,1
+user_93,Sales,Specialist,Remote,Oracle,23,0,1,0.439,0.454,24,15,7,2,0
+user_94,Engineering,Engineer,Remote,Jira,15,1,0,0.349,0.847,104,19,6,5,1
+user_95,HR,Analyst,SF,Workday,45,1,1,0.421,0.301,157,4,3,6,1
+user_96,Sales,Specialist,TX,Jira,31,1,1,0.895,0.215,83,16,8,10,1
+user_97,HR,Specialist,Remote,SAP,42,1,1,0.906,0.69,147,11,10,4,0
+user_98,Engineering,Engineer,NY,Jira,20,1,0,0.627,0.889,163,8,9,8,1
+user_99,Finance,Manager,TX,SAP,40,1,1,0.456,0.179,122,12,3,0,1
+user_100,HR,Admin,Remote,Workday,37,0,1,0.847,0.439,108,16,6,9,0
+user_101,Sales,Engineer,SF,Oracle,34,1,0,0.919,0.31,135,3,5,9,1
+user_102,Engineering,Manager,TX,SAP,32,1,0,0.518,0.615,5,11,1,6,1
+user_103,Engineering,Manager,Remote,SAP,11,1,1,0.553,0.951,92,11,10,4,0
+user_104,HR,Admin,TX,Jira,30,1,0,0.492,0.85,57,5,2,8,1
+user_105,Finance,Admin,Remote,GitHub,22,0,1,0.427,0.818,16,10,8,6,0
+user_106,HR,Specialist,SF,Jira,7,0,0,0.294,0.207,89,11,0,11,0
+user_107,Sales,Engineer,TX,Workday,12,0,1,0.301,0.23,61,0,8,1,0
+user_108,Support,Manager,Remote,Oracle,9,1,1,0.926,0.106,132,5,10,5,0
+user_109,Sales,Admin,Remote,Workday,9,1,0,0.335,0.915,160,19,10,0,1
+user_110,Engineering,Admin,TX,Jira,45,1,1,0.627,0.952,63,2,6,7,0
+user_111,HR,Analyst,TX,Oracle,10,0,0,0.688,0.958,161,18,3,0,0
+user_112,Engineering,Analyst,SF,GitHub,36,1,0,0.896,0.965,134,2,3,12,0
+user_113,Sales,Engineer,Remote,GitHub,19,0,1,0.356,0.334,115,3,0,2,0
+user_114,Finance,Manager,NY,SAP,15,0,1,0.292,0.925,154,7,8,12,0
+user_115,Engineering,Analyst,SF,Workday,23,0,1,0.564,0.85,6,12,7,9,0
+user_116,Sales,Specialist,SF,Jira,44,0,1,0.3,0.658,14,14,7,1,0
+user_117,Engineering,Manager,TX,Oracle,32,0,1,0.406,0.343,11,0,3,3,1
+user_118,Support,Admin,TX,GitHub,33,0,1,0.558,0.29,138,15,3,1,1
+user_119,Engineering,Admin,NY,Jira,11,1,1,0.642,0.803,179,4,4,8,1
+user_120,Sales,Specialist,TX,Workday,27,0,0,0.473,0.27,82,16,7,11,1
+user_121,Engineering,Manager,TX,Oracle,37,0,0,0.49,0.301,92,13,1,3,0
+user_122,Engineering,Admin,Remote,Workday,11,1,0,0.44,0.799,170,17,5,6,1
+user_123,Support,Engineer,NY,Jira,30,1,0,0.66,0.208,176,0,2,1,0
+user_124,Support,Manager,Remote,Jira,38,0,0,0.455,0.185,129,7,7,6,0
+user_125,HR,Admin,SF,Jira,15,0,1,0.798,0.907,18,3,8,9,0
+user_126,HR,Specialist,Remote,Oracle,45,0,0,0.604,0.585,78,13,3,11,0
+user_127,Sales,Admin,TX,Workday,30,0,1,0.34,0.781,10,16,5,9,0
+user_128,Sales,Engineer,SF,Jira,11,0,0,0.497,0.555,107,7,10,8,1
+user_129,HR,Engineer,Remote,SAP,11,1,1,0.718,0.238,145,8,2,10,1
+user_130,Engineering,Manager,Remote,Workday,7,0,0,0.692,0.838,83,9,4,12,0
+user_131,Sales,Admin,NY,Jira,31,1,1,0.241,0.962,127,4,3,6,0
+user_132,Sales,Manager,TX,Jira,11,1,0,0.776,0.72,39,10,4,7,1
+user_133,HR,Specialist,TX,Workday,10,0,0,0.501,0.62,28,5,4,4,1
+user_134,Finance,Admin,Remote,GitHub,44,1,0,0.674,0.707,108,14,1,3,0
+user_135,Support,Analyst,SF,Oracle,27,1,0,0.414,0.784,50,0,0,5,0
+user_136,Support,Manager,NY,SAP,39,1,0,0.473,0.34,61,3,0,4,1
+user_137,Engineering,Analyst,NY,SAP,14,1,1,0.383,0.53,91,17,0,10,0
+user_138,Finance,Admin,NY,Jira,42,0,0,0.212,0.68,176,3,5,0,0
+user_139,HR,Manager,TX,GitHub,38,1,1,0.939,0.624,116,10,2,2,1
+user_140,Support,Specialist,SF,Jira,6,0,0,0.778,0.307,66,9,4,12,1
+user_141,Finance,Engineer,Remote,Jira,18,1,0,0.908,0.349,80,3,5,0,1
+user_142,Support,Admin,NY,SAP,13,0,0,0.607,0.37,168,11,2,9,1
+user_143,Finance,Admin,NY,GitHub,32,1,1,0.349,0.6,166,12,10,9,1
+user_144,Sales,Analyst,Remote,SAP,37,0,1,0.307,0.391,92,11,5,3,0
+user_145,Finance,Engineer,SF,Jira,10,1,1,0.447,0.512,49,11,4,7,1
+user_146,Finance,Admin,TX,Jira,14,0,1,0.232,0.596,145,6,7,11,0
+user_147,Engineering,Specialist,SF,Workday,32,1,0,0.697,0.143,56,12,6,0,0
+user_148,HR,Engineer,TX,Jira,25,0,0,0.37,0.579,74,8,8,8,1
+user_149,Engineering,Specialist,SF,Workday,16,1,0,0.405,0.246,125,6,5,5,1
+user_150,Support,Analyst,TX,Oracle,15,0,0,0.583,0.236,95,7,7,1,0
+user_151,HR,Specialist,NY,Jira,38,1,0,0.266,0.79,40,14,8,10,0
+user_152,Finance,Specialist,Remote,SAP,34,0,0,0.48,0.77,63,7,2,6,0
+user_153,Finance,Manager,SF,SAP,32,0,0,0.662,0.683,75,14,10,2,0
+user_154,Sales,Engineer,TX,Oracle,10,0,1,0.386,0.863,21,6,2,6,0
+user_155,Finance,Manager,SF,Workday,33,0,0,0.649,0.984,130,10,5,7,1
+user_156,Support,Specialist,SF,SAP,27,0,1,0.23,0.481,161,7,9,3,1
+user_157,Support,Specialist,NY,Jira,35,1,1,0.624,0.121,53,7,3,4,0
+user_158,HR,Engineer,SF,GitHub,23,0,0,0.579,0.684,147,8,1,8,0
+user_159,Engineering,Manager,NY,SAP,26,0,1,0.276,0.32,40,17,6,11,1
+user_160,Engineering,Manager,SF,Oracle,23,1,0,0.624,0.549,144,8,1,2,1
+user_161,Support,Manager,NY,GitHub,17,0,0,0.205,0.95,140,0,2,0,1
+user_162,Support,Analyst,TX,Jira,13,1,0,0.657,0.492,133,15,9,4,1
+user_163,Engineering,Specialist,NY,Oracle,12,0,0,0.675,0.466,137,9,6,3,0
+user_164,Finance,Admin,SF,Oracle,14,1,1,0.285,0.949,149,17,9,10,0
+user_165,Support,Admin,Remote,Workday,33,1,0,0.926,0.484,125,17,10,0,1
+user_166,Sales,Admin,SF,GitHub,5,0,1,0.58,0.484,37,8,3,6,1
+user_167,Support,Engineer,TX,SAP,23,0,0,0.502,0.773,108,4,1,12,0
+user_168,Sales,Specialist,SF,SAP,31,0,0,0.745,0.733,157,19,0,6,1
+user_169,Support,Specialist,SF,Oracle,33,1,0,0.652,0.283,47,1,7,8,0
+user_170,Sales,Manager,NY,Workday,9,0,1,0.255,0.543,7,10,0,0,1
+user_171,Engineering,Specialist,SF,Jira,23,1,1,0.93,0.777,56,9,0,3,1
+user_172,Engineering,Specialist,Remote,Workday,25,0,0,0.872,0.568,142,9,0,10,0
+user_173,Finance,Engineer,SF,Oracle,6,0,1,0.378,0.408,99,5,0,11,0
+user_174,Sales,Manager,Remote,GitHub,10,0,1,0.486,0.76,146,17,10,4,1
+user_175,Engineering,Analyst,TX,Jira,31,0,1,0.388,0.981,18,4,6,12,0
+user_176,Engineering,Analyst,Remote,Workday,41,1,0,0.591,0.809,117,8,8,9,1
+user_177,Engineering,Analyst,SF,Oracle,10,0,1,0.889,0.143,93,16,10,12,0
+user_178,Engineering,Specialist,TX,Workday,37,0,0,0.642,0.721,90,2,1,9,0
+user_179,Engineering,Engineer,Remote,Workday,36,0,0,0.659,0.556,66,2,1,8,0
+user_180,Engineering,Engineer,NY,SAP,38,0,1,0.247,0.168,59,17,1,12,0
+user_181,Support,Manager,SF,Workday,44,0,1,0.466,0.733,98,12,5,8,1
+user_182,Sales,Analyst,NY,SAP,20,0,0,0.916,0.476,61,13,2,12,0
+user_183,Sales,Manager,TX,Jira,5,1,1,0.375,0.213,84,18,5,4,1
+user_184,Support,Engineer,Remote,Oracle,5,1,1,0.821,0.644,76,4,2,0,0
+user_185,Finance,Specialist,Remote,Workday,17,0,1,0.866,0.285,52,1,8,7,1
+user_186,Finance,Manager,NY,GitHub,6,0,1,0.895,0.704,118,13,7,2,1
+user_187,HR,Manager,Remote,Jira,32,0,0,0.308,0.41,67,18,8,12,0
+user_188,Finance,Analyst,SF,Jira,23,0,1,0.598,0.686,108,1,4,11,1
+user_189,Support,Analyst,NY,GitHub,5,0,1,0.621,0.42,102,8,2,6,1
+user_190,Finance,Specialist,Remote,GitHub,21,1,1,0.56,0.325,89,10,1,12,0
+user_191,HR,Engineer,Remote,GitHub,11,0,0,0.636,0.892,92,19,4,9,1
+user_192,Finance,Manager,TX,GitHub,31,0,0,0.715,0.682,120,4,10,0,0
+user_193,Finance,Analyst,NY,SAP,15,1,0,0.262,0.604,74,0,8,2,1
+user_194,Sales,Manager,Remote,Workday,21,1,0,0.539,0.105,52,1,2,10,0
+user_195,HR,Specialist,TX,Oracle,33,0,0,0.699,0.924,56,12,4,11,1
+user_196,Finance,Manager,NY,Workday,43,1,1,0.517,0.114,77,15,2,7,0
+user_197,Sales,Admin,SF,GitHub,12,1,1,0.858,0.159,83,6,4,0,1
+user_198,Finance,Manager,SF,Workday,12,0,0,0.463,0.195,3,1,7,10,1
+user_199,Engineering,Analyst,TX,GitHub,17,0,0,0.841,0.57,107,6,1,8,1
+user_200,Support,Engineer,Remote,Workday,38,1,0,0.692,0.622,11,13,0,6,0
+user_201,HR,Manager,Remote,Jira,29,0,1,0.596,0.192,29,9,3,11,1
+user_202,Finance,Admin,NY,GitHub,13,0,1,0.754,0.572,14,20,7,2,0
+user_203,Sales,Analyst,TX,GitHub,20,0,0,0.428,0.802,130,15,0,9,1
+user_204,Support,Engineer,Remote,Jira,25,1,0,0.502,0.742,44,18,1,4,1
+user_205,Finance,Analyst,Remote,Oracle,42,0,1,0.883,0.144,131,16,7,12,0
+user_206,Engineering,Manager,SF,Workday,43,1,1,0.456,0.345,14,7,2,0,1
+user_207,Finance,Engineer,Remote,Jira,41,1,1,0.406,0.145,68,8,4,8,0
+user_208,Finance,Analyst,Remote,GitHub,33,1,0,0.913,0.948,23,3,7,1,1
+user_209,Sales,Admin,Remote,Oracle,31,1,1,0.266,0.43,133,3,0,11,1
+user_210,Sales,Engineer,TX,Oracle,45,0,0,0.657,0.377,9,12,6,4,1
+user_211,Sales,Analyst,NY,GitHub,31,1,0,0.465,0.444,138,20,5,0,0
+user_212,Finance,Analyst,TX,GitHub,33,0,0,0.837,0.182,37,11,9,1,0
+user_213,Finance,Specialist,TX,Jira,32,0,1,0.948,0.61,39,20,7,12,0
+user_214,Finance,Admin,TX,GitHub,42,0,1,0.346,0.928,156,20,7,12,1
+user_215,Support,Manager,TX,GitHub,10,1,0,0.759,0.82,138,3,0,2,1
+user_216,Finance,Engineer,Remote,SAP,8,0,1,0.41,0.741,9,14,6,0,0
+user_217,Sales,Analyst,SF,Oracle,20,1,1,0.746,0.7,129,12,9,1,1
+user_218,Engineering,Analyst,SF,Oracle,9,0,0,0.871,0.645,82,13,2,7,1
+user_219,HR,Engineer,Remote,Workday,15,1,0,0.545,0.395,12,11,10,11,1
+user_220,Support,Engineer,NY,SAP,18,1,0,0.491,0.478,178,11,10,1,0
+user_221,Engineering,Analyst,SF,SAP,43,0,1,0.638,0.214,26,19,3,5,0
+user_222,Support,Admin,SF,GitHub,44,1,1,0.835,0.6,30,2,9,7,0
+user_223,Sales,Specialist,NY,Workday,45,0,0,0.49,0.404,24,15,1,3,1
+user_224,Engineering,Engineer,NY,Oracle,12,0,1,0.736,0.896,110,16,7,6,1
+user_225,Engineering,Engineer,TX,GitHub,5,1,0,0.637,0.83,29,14,10,0,0
+user_226,HR,Specialist,Remote,Jira,25,0,0,0.345,0.534,104,10,3,11,0
+user_227,Engineering,Engineer,SF,Workday,36,0,1,0.225,0.631,139,11,10,12,0
+user_228,HR,Specialist,NY,Jira,41,0,0,0.909,0.871,107,19,4,4,0
+user_229,HR,Analyst,SF,SAP,27,1,0,0.66,0.313,174,12,6,2,0
+user_230,HR,Admin,TX,Jira,18,0,0,0.204,0.257,50,13,4,1,0
+user_231,Finance,Manager,Remote,Oracle,30,0,0,0.212,0.431,14,2,5,5,0
+user_232,Support,Specialist,TX,Oracle,16,0,1,0.846,0.143,105,0,10,12,0
+user_233,HR,Engineer,Remote,Oracle,21,0,1,0.652,0.637,115,15,1,1,1
+user_234,Support,Analyst,TX,Jira,14,0,0,0.281,0.965,19,15,6,10,0
+user_235,Sales,Manager,SF,Workday,34,0,1,0.27,0.428,79,13,4,12,1
+user_236,HR,Specialist,TX,SAP,23,1,0,0.37,0.827,42,1,8,1,1
+user_237,Sales,Analyst,TX,SAP,10,0,1,0.483,0.538,173,10,2,11,0
+user_238,Engineering,Engineer,TX,Workday,23,0,1,0.554,0.94,110,15,0,5,1
+user_239,Support,Engineer,NY,Jira,29,1,1,0.909,0.847,4,0,9,0,1
+user_240,Sales,Manager,TX,Jira,34,1,1,0.776,0.664,76,1,0,11,1
+user_241,HR,Analyst,TX,Jira,36,0,1,0.677,0.173,4,15,3,0,0
+user_242,HR,Admin,SF,Oracle,27,0,0,0.872,0.772,48,14,9,10,0
+user_243,Engineering,Analyst,TX,Jira,32,0,1,0.699,0.684,174,2,2,7,0
+user_244,HR,Specialist,SF,Jira,43,0,1,0.343,0.209,19,9,2,10,0
+user_245,Sales,Admin,TX,SAP,36,0,0,0.201,0.139,140,2,2,10,1
+user_246,Sales,Engineer,Remote,Jira,10,1,0,0.47,0.407,129,1,7,10,1
+user_247,Support,Engineer,TX,Workday,11,0,1,0.788,0.686,168,12,4,1,0
+user_248,Sales,Manager,NY,Oracle,9,0,1,0.333,0.246,101,6,8,10,1
+user_249,Finance,Engineer,Remote,SAP,9,1,0,0.73,0.706,71,15,2,10,1
+user_250,Finance,Specialist,NY,Workday,6,0,1,0.894,0.393,29,7,10,1,1
+user_251,Sales,Admin,Remote,SAP,14,1,0,0.593,0.83,39,7,4,4,0
+user_252,Finance,Analyst,NY,Workday,18,1,1,0.304,0.558,123,16,2,12,1
+user_253,Engineering,Manager,Remote,Jira,5,1,0,0.832,0.534,33,13,6,8,1
+user_254,Finance,Manager,Remote,GitHub,30,0,1,0.233,0.42,36,12,3,2,0
+user_255,Sales,Engineer,TX,GitHub,20,0,0,0.911,0.638,133,5,0,0,1
+user_256,Sales,Admin,NY,GitHub,44,1,0,0.206,0.565,173,11,2,8,0
+user_257,Support,Analyst,SF,Workday,24,0,1,0.368,0.559,64,6,8,0,0
+user_258,Support,Analyst,NY,GitHub,29,0,0,0.66,0.76,124,2,5,8,1
+user_259,Engineering,Specialist,SF,Jira,17,1,1,0.392,0.223,180,18,7,8,1
+user_260,HR,Specialist,SF,Jira,17,1,0,0.907,0.113,12,9,0,2,1
+user_261,Finance,Manager,NY,SAP,16,1,0,0.344,0.166,5,9,4,8,1
+user_262,Engineering,Engineer,Remote,SAP,16,0,1,0.726,0.782,173,12,9,8,1
+user_263,HR,Admin,SF,Workday,30,1,0,0.376,0.771,166,5,2,11,0
+user_264,Engineering,Analyst,SF,Workday,23,1,0,0.503,0.529,62,3,7,0,0
+user_265,Engineering,Admin,SF,Jira,17,0,1,0.487,0.537,114,9,0,1,0
+user_266,Support,Manager,Remote,Jira,24,1,0,0.649,0.758,138,10,1,1,1
+user_267,HR,Admin,TX,Workday,34,0,1,0.374,0.662,152,0,1,9,1
+user_268,HR,Engineer,NY,Workday,9,1,0,0.804,0.481,104,19,10,3,1
+user_269,Finance,Admin,TX,Workday,37,1,1,0.515,0.379,14,12,2,8,1
+user_270,Engineering,Admin,Remote,Jira,35,1,0,0.322,0.156,58,8,10,2,0
+user_271,Sales,Manager,NY,Oracle,26,0,1,0.793,0.148,88,0,9,8,1
+user_272,Support,Engineer,Remote,SAP,36,1,0,0.853,0.392,16,0,5,12,0
+user_273,Sales,Specialist,Remote,Oracle,35,1,1,0.621,0.964,104,2,5,11,1
+user_274,Engineering,Manager,SF,Jira,10,0,0,0.258,0.425,28,14,9,12,1
+user_275,Finance,Analyst,Remote,Jira,31,0,1,0.284,0.495,80,14,0,5,1
+user_276,Finance,Engineer,NY,Jira,20,1,0,0.454,0.736,127,15,0,11,0
+user_277,HR,Analyst,SF,Jira,30,1,1,0.389,0.244,144,0,6,10,1
+user_278,Engineering,Analyst,TX,Oracle,26,0,1,0.686,0.393,40,17,6,12,0
+user_279,Finance,Specialist,TX,Oracle,9,1,1,0.633,0.936,69,15,3,4,1
+user_280,Sales,Admin,NY,Workday,41,1,0,0.923,0.404,127,14,9,5,1
+user_281,HR,Specialist,TX,Jira,6,0,1,0.328,0.551,163,16,10,7,1
+user_282,HR,Engineer,SF,Jira,23,0,1,0.861,0.771,26,17,9,5,0
+user_283,Support,Manager,NY,GitHub,27,0,1,0.234,0.275,141,2,1,2,1
+user_284,Finance,Manager,NY,GitHub,23,1,0,0.638,0.332,142,13,6,12,0
+user_285,Support,Manager,Remote,GitHub,27,0,0,0.748,0.95,102,19,6,8,1
+user_286,Support,Specialist,NY,GitHub,36,1,1,0.949,0.972,56,7,8,10,0
+user_287,Finance,Engineer,TX,SAP,33,0,0,0.625,0.497,119,20,4,8,1
+user_288,HR,Analyst,TX,Oracle,12,1,1,0.333,0.904,60,10,0,9,0
+user_289,Finance,Admin,Remote,Jira,28,0,1,0.931,0.316,81,18,3,2,0
+user_290,Engineering,Analyst,TX,Oracle,11,1,1,0.519,0.54,7,2,2,11,1
+user_291,HR,Engineer,TX,Workday,38,1,1,0.464,0.245,162,3,8,10,1
+user_292,Engineering,Manager,Remote,Workday,37,1,1,0.306,0.931,67,2,7,6,0
+user_293,HR,Engineer,TX,SAP,18,1,0,0.72,0.285,103,9,4,8,0
+user_294,Sales,Analyst,Remote,Oracle,21,1,0,0.309,0.49,1,9,9,10,1
+user_295,Sales,Analyst,Remote,Jira,19,1,1,0.364,0.39,117,3,1,1,0
+user_296,HR,Analyst,SF,Workday,12,0,0,0.7,0.638,29,2,2,5,1
+user_297,Engineering,Specialist,NY,SAP,42,0,1,0.579,0.97,107,3,4,7,0
+user_298,HR,Analyst,Remote,SAP,39,0,0,0.377,0.617,128,20,0,4,1
+user_299,HR,Manager,Remote,SAP,29,0,0,0.616,0.958,97,10,8,9,0
+user_300,Support,Analyst,TX,Workday,38,1,0,0.9,0.169,161,16,3,3,0
+user_301,Support,Specialist,TX,Workday,31,1,0,0.313,0.811,53,5,4,1,0
+user_302,HR,Engineer,TX,SAP,14,1,1,0.633,0.36,89,19,5,0,1
+user_303,Engineering,Specialist,Remote,Workday,25,0,0,0.624,0.135,0,11,7,3,0
+user_304,Sales,Manager,NY,Jira,8,0,1,0.232,0.132,163,15,1,7,1
+user_305,Sales,Specialist,TX,SAP,13,1,1,0.441,0.398,165,17,4,6,0
+user_306,Support,Specialist,NY,Workday,42,1,0,0.31,0.147,51,5,0,9,0
+user_307,Sales,Analyst,NY,SAP,43,0,0,0.567,0.56,26,20,3,2,0
+user_308,HR,Specialist,Remote,Oracle,37,0,1,0.826,0.203,122,11,3,10,1
+user_309,Engineering,Analyst,TX,GitHub,9,1,1,0.335,0.729,116,3,0,9,0
+user_310,Finance,Admin,Remote,Oracle,23,0,0,0.594,0.881,116,9,3,1,0
+user_311,Support,Manager,TX,GitHub,26,1,0,0.504,0.526,21,5,2,8,0
+user_312,Engineering,Specialist,Remote,Oracle,14,0,1,0.674,0.739,12,19,1,4,0
+user_313,HR,Manager,NY,GitHub,17,1,0,0.313,0.779,150,12,0,3,1
+user_314,HR,Manager,NY,GitHub,33,0,1,0.747,0.163,143,1,2,9,0
+user_315,Support,Analyst,SF,GitHub,9,1,0,0.369,0.976,141,1,3,3,0
+user_316,Finance,Engineer,SF,Oracle,19,0,1,0.91,0.695,163,9,4,8,0
+user_317,Engineering,Specialist,Remote,SAP,36,1,0,0.724,0.699,165,4,10,9,0
+user_318,HR,Analyst,TX,SAP,36,1,0,0.54,0.988,166,4,8,3,1
+user_319,HR,Admin,NY,Jira,26,0,0,0.866,0.862,135,11,0,3,1
+user_320,Sales,Admin,Remote,GitHub,34,1,0,0.391,0.213,40,5,10,2,1
+user_321,Support,Analyst,NY,GitHub,21,0,1,0.255,0.232,135,16,7,12,0
+user_322,HR,Specialist,SF,Workday,20,0,1,0.703,0.204,20,14,9,9,1
+user_323,HR,Admin,TX,Oracle,6,1,0,0.856,0.86,179,11,5,2,0
+user_324,Support,Manager,Remote,Workday,24,0,0,0.486,0.615,121,4,10,1,0
+user_325,Finance,Manager,SF,SAP,31,0,0,0.832,0.616,148,7,9,0,0
+user_326,Engineering,Manager,SF,GitHub,9,1,0,0.89,0.674,96,4,0,2,0
+user_327,Engineering,Analyst,TX,GitHub,13,0,0,0.304,0.77,16,10,7,10,0
+user_328,Engineering,Specialist,TX,Oracle,42,1,0,0.45,0.644,130,20,5,9,1
+user_329,Support,Admin,NY,GitHub,17,1,1,0.263,0.926,0,12,2,2,1
+user_330,Engineering,Specialist,Remote,Jira,45,1,1,0.461,0.49,88,17,2,9,0
+user_331,Engineering,Specialist,TX,Jira,29,0,0,0.228,0.645,130,20,10,5,0
+user_332,HR,Specialist,TX,Workday,25,0,0,0.896,0.777,32,11,3,8,0
+user_333,Finance,Admin,Remote,Workday,11,0,1,0.398,0.468,132,12,4,5,1
+user_334,Engineering,Analyst,SF,GitHub,29,0,1,0.641,0.119,89,8,3,11,0
+user_335,Engineering,Admin,SF,SAP,39,0,1,0.354,0.284,143,13,1,2,0
+user_336,Sales,Admin,TX,Oracle,35,1,0,0.742,0.505,48,16,9,3,1
+user_337,Support,Admin,SF,GitHub,23,1,0,0.563,0.389,80,15,3,9,1
+user_338,Engineering,Admin,Remote,GitHub,7,1,1,0.437,0.757,11,20,3,9,1
+user_339,Finance,Engineer,Remote,SAP,38,0,0,0.579,0.79,141,10,4,3,1
+user_340,Finance,Manager,TX,Jira,25,1,0,0.869,0.249,110,18,7,7,0
+user_341,Support,Analyst,TX,Workday,32,0,0,0.369,0.915,91,3,0,10,0
+user_342,Support,Admin,Remote,GitHub,32,1,0,0.329,0.562,93,7,1,4,1
+user_343,HR,Analyst,NY,Oracle,25,1,1,0.91,0.418,70,11,8,11,1
+user_344,Finance,Engineer,TX,SAP,7,0,1,0.947,0.662,8,19,8,2,1
+user_345,Engineering,Manager,NY,Jira,28,0,0,0.945,0.873,148,6,1,1,1
+user_346,Sales,Engineer,TX,Jira,5,0,0,0.809,0.625,57,10,3,11,0
+user_347,HR,Engineer,SF,Workday,9,0,0,0.602,0.338,66,18,5,7,0
+user_348,Support,Analyst,Remote,Workday,17,0,1,0.628,0.188,151,11,3,0,0
+user_349,Engineering,Admin,SF,Jira,38,1,0,0.352,0.29,38,15,6,10,1
+user_350,HR,Specialist,Remote,Oracle,36,0,0,0.249,0.479,117,12,2,0,1
+user_351,Support,Manager,TX,GitHub,27,0,0,0.314,0.173,159,20,8,11,0
+user_352,HR,Engineer,NY,Jira,8,0,1,0.755,0.562,25,1,8,2,0
+user_353,Finance,Analyst,NY,SAP,18,0,0,0.306,0.154,20,1,5,9,0
+user_354,HR,Analyst,NY,Workday,21,1,1,0.355,0.114,36,12,5,11,0
+user_355,Engineering,Specialist,SF,Workday,25,0,0,0.426,0.124,137,3,0,12,1
+user_356,Support,Engineer,SF,Oracle,37,0,0,0.542,0.622,3,20,6,0,0
+user_357,Finance,Analyst,NY,Jira,44,1,1,0.305,0.313,10,7,9,8,0
+user_358,Engineering,Engineer,Remote,SAP,30,1,1,0.934,0.469,66,15,9,8,1
+user_359,Finance,Manager,TX,GitHub,12,0,1,0.314,0.458,172,14,2,7,0
+user_360,Support,Admin,SF,SAP,25,1,0,0.581,0.581,63,3,8,4,1
+user_361,Support,Manager,TX,GitHub,30,1,0,0.379,0.256,74,16,10,2,1
+user_362,Support,Analyst,Remote,SAP,11,1,1,0.397,0.212,48,7,9,6,0
+user_363,Finance,Specialist,TX,Workday,40,1,1,0.25,0.466,20,19,0,0,1
+user_364,Sales,Specialist,Remote,Jira,42,0,0,0.269,0.529,54,3,0,0,1
+user_365,Sales,Specialist,Remote,Jira,20,0,1,0.746,0.235,162,1,10,10,0
+user_366,Engineering,Admin,SF,Workday,6,0,1,0.337,0.631,24,15,6,1,1
+user_367,Support,Admin,NY,Jira,40,1,1,0.544,0.26,4,7,6,2,0
+user_368,Support,Manager,TX,GitHub,12,1,1,0.734,0.15,131,3,10,7,0
+user_369,Engineering,Admin,Remote,Jira,44,0,0,0.709,0.606,129,0,6,3,0
+user_370,Engineering,Analyst,TX,Jira,9,0,1,0.464,0.75,153,9,7,9,1
+user_371,Sales,Analyst,SF,SAP,40,0,1,0.323,0.212,52,11,4,9,1
+user_372,Engineering,Specialist,TX,SAP,26,1,1,0.694,0.61,47,10,7,3,1
+user_373,Support,Admin,NY,Workday,9,0,0,0.66,0.131,66,11,2,5,1
+user_374,Support,Admin,SF,SAP,35,1,1,0.723,0.636,179,5,0,9,0
+user_375,Engineering,Engineer,TX,Jira,27,1,0,0.661,0.492,3,20,4,9,0
+user_376,Finance,Analyst,TX,SAP,40,1,1,0.242,0.781,180,8,1,5,0
+user_377,Support,Manager,SF,Jira,38,0,0,0.432,0.704,141,20,1,12,1
+user_378,Finance,Manager,SF,GitHub,5,0,0,0.556,0.837,154,13,10,5,1
+user_379,HR,Analyst,SF,Workday,21,0,0,0.562,0.392,158,5,2,9,1
+user_380,Support,Specialist,NY,SAP,40,1,1,0.627,0.656,9,14,5,12,0
+user_381,HR,Engineer,TX,GitHub,9,1,0,0.495,0.389,121,0,4,7,0
+user_382,HR,Engineer,Remote,GitHub,5,1,1,0.947,0.102,71,8,1,6,1
+user_383,HR,Specialist,Remote,Jira,28,1,1,0.218,0.689,67,9,8,1,1
+user_384,Support,Analyst,NY,Workday,30,0,0,0.21,0.468,65,15,10,5,0
+user_385,Sales,Manager,SF,GitHub,28,0,1,0.596,0.321,53,13,7,6,0
+user_386,Sales,Analyst,TX,Oracle,6,1,0,0.785,0.467,143,0,4,1,0
+user_387,Support,Engineer,TX,GitHub,30,0,1,0.678,0.11,5,9,1,9,1
+user_388,Support,Admin,Remote,Jira,9,0,0,0.848,0.322,75,2,2,1,1
+user_389,Finance,Specialist,Remote,SAP,30,0,0,0.833,0.958,115,2,4,5,0
+user_390,HR,Engineer,Remote,Oracle,17,0,0,0.35,0.836,149,15,0,1,0
+user_391,Finance,Specialist,Remote,Workday,10,0,1,0.901,0.23,121,13,10,1,0
+user_392,HR,Specialist,TX,Workday,43,0,1,0.776,0.793,100,15,9,1,1
+user_393,Finance,Manager,NY,Jira,13,1,1,0.619,0.492,69,8,1,1,0
+user_394,Support,Admin,NY,SAP,40,1,1,0.464,0.931,57,9,1,4,0
+user_395,Finance,Specialist,SF,GitHub,21,0,0,0.642,0.25,67,4,10,10,0
+user_396,Support,Engineer,Remote,SAP,25,0,0,0.244,0.72,111,10,2,9,1
+user_397,Finance,Manager,NY,Oracle,37,1,0,0.587,0.374,96,7,1,3,1
+user_398,Finance,Analyst,TX,Workday,37,0,0,0.59,0.472,86,13,7,3,1
+user_399,Finance,Engineer,Remote,Oracle,33,0,1,0.389,0.125,82,5,5,8,1
+user_400,Sales,Manager,TX,GitHub,36,0,0,0.401,0.972,4,12,9,5,1
+user_401,Engineering,Analyst,TX,Workday,21,0,0,0.376,0.455,59,20,1,1,0
+user_402,Engineering,Manager,Remote,Jira,42,0,0,0.602,0.479,35,10,8,4,1
+user_403,Engineering,Engineer,TX,Oracle,27,1,1,0.579,0.597,137,3,3,7,1
+user_404,Sales,Admin,NY,Workday,27,1,0,0.298,0.16,141,6,7,6,0
+user_405,Finance,Manager,NY,SAP,17,0,0,0.833,0.219,75,13,7,9,0
+user_406,Sales,Engineer,NY,Oracle,10,0,0,0.408,0.631,124,3,10,5,0
+user_407,HR,Specialist,NY,Jira,6,1,1,0.214,0.927,56,4,8,11,0
+user_408,Finance,Engineer,Remote,Workday,39,0,1,0.657,0.275,47,4,4,8,1
+user_409,HR,Engineer,NY,Oracle,18,1,1,0.671,0.471,111,6,8,1,1
+user_410,Support,Specialist,SF,Workday,37,0,1,0.253,0.326,58,6,4,4,1
+user_411,Finance,Specialist,TX,Jira,17,0,0,0.338,0.716,156,15,10,6,1
+user_412,HR,Analyst,SF,Workday,34,1,1,0.844,0.662,69,5,10,11,1
+user_413,Support,Manager,TX,Oracle,13,0,0,0.708,0.821,176,13,3,0,0
+user_414,Engineering,Engineer,SF,GitHub,7,0,1,0.884,0.875,151,16,1,2,1
+user_415,Engineering,Engineer,TX,Workday,12,0,0,0.438,0.769,150,8,8,7,0
+user_416,Sales,Specialist,TX,Workday,15,1,1,0.566,0.791,66,15,1,4,0
+user_417,Engineering,Engineer,Remote,Workday,7,1,0,0.248,0.888,176,1,7,6,1
+user_418,Support,Specialist,NY,SAP,6,1,1,0.883,0.68,131,19,5,4,1
+user_419,Sales,Admin,SF,GitHub,39,0,1,0.765,0.188,29,12,7,3,1
+user_420,Sales,Manager,SF,Jira,38,1,0,0.395,0.293,120,18,4,0,1
+user_421,Sales,Manager,NY,SAP,21,1,0,0.348,0.931,71,12,4,8,0
+user_422,Engineering,Engineer,TX,SAP,9,0,0,0.925,0.629,30,15,7,1,0
+user_423,Support,Analyst,SF,GitHub,25,0,0,0.408,0.693,103,10,4,2,1
+user_424,HR,Specialist,Remote,Oracle,19,0,0,0.753,0.734,71,15,8,7,1
+user_425,Finance,Manager,NY,GitHub,30,1,1,0.799,0.816,30,4,4,7,0
+user_426,HR,Admin,Remote,Oracle,5,0,1,0.587,0.486,162,7,8,3,0
+user_427,Sales,Admin,TX,SAP,35,0,0,0.841,0.785,128,4,0,5,0
+user_428,Finance,Admin,SF,GitHub,19,1,1,0.268,0.175,29,16,10,2,1
+user_429,Sales,Engineer,SF,SAP,32,1,0,0.472,0.839,125,8,9,0,1
+user_430,HR,Analyst,TX,GitHub,30,0,1,0.776,0.397,18,18,6,10,1
+user_431,Engineering,Manager,TX,SAP,39,0,1,0.376,0.17,96,9,4,2,0
+user_432,Support,Specialist,SF,GitHub,18,1,1,0.942,0.536,111,5,8,12,1
+user_433,HR,Admin,NY,GitHub,41,0,0,0.876,0.388,130,14,2,5,1
+user_434,Sales,Admin,SF,Jira,44,1,0,0.228,0.254,76,12,4,11,1
+user_435,Finance,Admin,Remote,Jira,29,1,1,0.565,0.267,115,7,9,10,1
+user_436,Sales,Engineer,NY,Jira,16,0,0,0.712,0.19,117,17,0,6,0
+user_437,Engineering,Manager,NY,Workday,38,1,0,0.422,0.492,28,6,4,10,1
+user_438,Finance,Analyst,SF,Oracle,40,0,0,0.427,0.943,136,8,6,2,0
+user_439,Engineering,Engineer,SF,SAP,26,1,1,0.652,0.666,24,6,0,8,1
+user_440,HR,Admin,Remote,SAP,39,0,0,0.949,0.674,177,11,4,2,1
+user_441,HR,Manager,NY,SAP,18,1,0,0.565,0.727,141,20,3,11,1
+user_442,Sales,Admin,Remote,SAP,12,0,0,0.41,0.225,101,5,9,5,0
+user_443,Engineering,Engineer,Remote,Workday,6,0,0,0.722,0.556,153,3,2,12,1
+user_444,Finance,Manager,Remote,Jira,8,0,0,0.503,0.123,169,17,9,12,0
+user_445,Sales,Analyst,SF,Workday,17,1,0,0.833,0.746,161,1,9,4,1
+user_446,Support,Engineer,Remote,Jira,33,0,0,0.294,0.385,101,20,0,10,0
+user_447,Support,Engineer,NY,Oracle,23,0,1,0.585,0.404,49,17,9,7,0
+user_448,Engineering,Admin,SF,GitHub,35,0,0,0.616,0.697,133,5,0,1,0
+user_449,Engineering,Admin,SF,Workday,27,1,1,0.865,0.513,141,13,0,1,1
+user_450,Engineering,Specialist,NY,GitHub,17,0,0,0.398,0.607,58,9,5,5,1
+user_451,Finance,Engineer,Remote,Oracle,24,1,1,0.442,0.89,88,8,9,1,0
+user_452,Sales,Analyst,Remote,Workday,10,0,0,0.23,0.112,11,14,6,11,1
+user_453,Finance,Manager,TX,Workday,19,0,1,0.729,0.639,92,1,0,6,0
+user_454,Engineering,Manager,TX,Oracle,41,1,0,0.825,0.975,98,13,5,9,1
+user_455,Support,Manager,SF,Jira,12,1,0,0.562,0.176,26,17,5,8,0
+user_456,Engineering,Admin,SF,SAP,6,0,1,0.773,0.825,154,11,0,0,1
+user_457,Sales,Admin,Remote,Workday,8,0,0,0.379,0.254,5,10,0,4,1
+user_458,HR,Admin,TX,Workday,17,0,0,0.393,0.858,101,5,4,2,1
+user_459,Engineering,Engineer,NY,Oracle,38,1,1,0.489,0.956,166,8,5,4,0
+user_460,HR,Admin,Remote,SAP,36,0,0,0.28,0.845,84,0,7,4,1
+user_461,Support,Engineer,TX,SAP,17,1,0,0.4,0.184,52,16,0,3,1
+user_462,Support,Specialist,Remote,SAP,11,1,0,0.578,0.585,152,9,8,9,1
+user_463,Sales,Specialist,NY,Jira,33,0,0,0.235,0.735,111,1,9,7,1
+user_464,HR,Analyst,TX,GitHub,24,1,0,0.601,0.764,172,18,9,4,1
+user_465,Support,Specialist,Remote,Jira,15,1,0,0.763,0.42,105,8,10,8,0
+user_466,HR,Manager,TX,Oracle,7,0,0,0.722,0.875,48,2,0,0,1
+user_467,Finance,Engineer,TX,Oracle,8,0,1,0.742,0.413,49,17,1,6,1
+user_468,Finance,Admin,TX,Jira,40,0,1,0.689,0.98,42,11,5,9,0
+user_469,HR,Engineer,Remote,Oracle,15,1,1,0.526,0.377,12,2,8,12,1
+user_470,Finance,Admin,Remote,Jira,33,0,0,0.926,0.377,100,14,0,10,1
+user_471,Finance,Specialist,TX,Workday,36,1,0,0.433,0.968,121,13,8,11,1
+user_472,HR,Manager,SF,GitHub,16,0,1,0.91,0.558,175,11,4,10,0
+user_473,HR,Manager,TX,Jira,24,0,1,0.61,0.853,171,8,4,12,0
+user_474,Finance,Analyst,SF,Workday,8,0,1,0.417,0.972,36,15,5,9,1
+user_475,Engineering,Analyst,TX,GitHub,19,0,0,0.589,0.479,145,12,10,5,1
+user_476,Engineering,Analyst,SF,GitHub,20,1,0,0.896,0.114,5,4,9,0,1
+user_477,Sales,Specialist,TX,Jira,26,0,0,0.346,0.846,63,8,4,11,1
+user_478,HR,Specialist,SF,SAP,36,0,0,0.289,0.46,95,14,9,12,0
+user_479,Support,Admin,Remote,GitHub,40,0,1,0.874,0.889,167,16,8,2,0
+user_480,HR,Manager,NY,Jira,15,1,0,0.54,0.557,156,7,4,11,0
+user_481,Finance,Manager,TX,SAP,10,0,1,0.784,0.102,24,5,1,2,1
+user_482,Support,Manager,NY,SAP,25,0,1,0.518,0.817,5,12,0,11,1
+user_483,Finance,Admin,SF,Workday,7,1,0,0.565,0.627,19,6,3,10,0
+user_484,HR,Manager,NY,GitHub,30,0,0,0.533,0.519,102,7,9,8,0
+user_485,Support,Analyst,Remote,SAP,40,1,0,0.618,0.588,138,1,0,12,0
+user_486,Engineering,Engineer,Remote,Oracle,35,0,1,0.712,0.579,26,6,7,1,0
+user_487,HR,Engineer,SF,SAP,25,1,1,0.687,0.757,116,6,9,7,0
+user_488,HR,Engineer,SF,Jira,22,0,1,0.234,0.452,113,12,2,10,1
+user_489,Sales,Admin,NY,Workday,21,1,1,0.3,0.745,43,8,6,11,1
+user_490,Engineering,Engineer,Remote,SAP,23,0,0,0.204,0.259,15,1,6,6,1
+user_491,HR,Admin,SF,Jira,40,1,1,0.778,0.936,32,19,4,2,1
+user_492,Sales,Analyst,NY,GitHub,22,0,1,0.344,0.652,121,13,6,11,1
+user_493,HR,Engineer,NY,Workday,16,1,0,0.508,0.283,39,7,1,2,0
+user_494,Finance,Specialist,TX,Jira,10,0,1,0.833,0.311,180,7,3,0,0
+user_495,Support,Specialist,TX,SAP,7,1,1,0.335,0.611,39,14,4,3,0
+user_496,Finance,Manager,NY,Workday,44,1,0,0.545,0.963,70,4,10,0,0
+user_497,HR,Admin,NY,Workday,8,1,0,0.619,0.142,70,4,3,3,0
+user_498,Sales,Specialist,Remote,SAP,42,0,1,0.276,0.308,155,12,1,3,1
+user_499,HR,Specialist,Remote,Oracle,30,0,1,0.931,0.284,130,11,7,5,0
+user_500,Engineering,Engineer,Remote,Oracle,16,1,0,0.641,0.601,21,20,10,12,1
+user_501,HR,Admin,Remote,SAP,28,0,1,0.567,0.196,59,17,2,9,1
+user_502,Sales,Engineer,TX,Workday,16,1,1,0.778,0.344,7,8,9,0,0
+user_503,Support,Admin,SF,GitHub,44,0,0,0.458,0.696,152,4,10,11,0
+user_504,Support,Engineer,TX,Workday,22,0,0,0.477,0.412,155,9,9,12,0
+user_505,Support,Analyst,Remote,Workday,31,1,1,0.539,0.327,120,4,3,7,1
+user_506,Sales,Analyst,TX,SAP,29,0,1,0.421,0.112,27,1,6,0,0
+user_507,HR,Manager,SF,Oracle,43,1,1,0.613,0.92,148,14,6,6,0
+user_508,Engineering,Manager,Remote,Jira,21,0,1,0.707,0.63,142,6,9,5,1
+user_509,Support,Analyst,TX,Jira,17,0,0,0.715,0.792,53,0,1,3,1
+user_510,Support,Analyst,Remote,GitHub,28,0,0,0.906,0.88,20,14,10,3,1
+user_511,Support,Admin,TX,SAP,35,0,0,0.379,0.543,64,6,3,4,0
+user_512,Support,Specialist,Remote,Oracle,43,1,0,0.776,0.508,150,4,7,9,0
+user_513,Engineering,Analyst,Remote,Workday,8,1,0,0.492,0.334,122,7,8,4,0
+user_514,Sales,Engineer,NY,GitHub,19,0,1,0.303,0.22,148,15,10,11,1
+user_515,Engineering,Specialist,NY,Jira,36,0,0,0.685,0.967,61,6,1,8,1
+user_516,Support,Specialist,TX,Workday,26,1,1,0.687,0.409,128,16,3,11,1
+user_517,Support,Admin,TX,GitHub,16,0,0,0.62,0.175,3,14,4,12,1
+user_518,Finance,Specialist,NY,Workday,33,1,1,0.243,0.694,42,15,1,11,1
+user_519,HR,Specialist,TX,SAP,17,1,0,0.251,0.618,54,4,1,0,0
+user_520,HR,Manager,TX,Workday,11,1,1,0.399,0.496,106,11,6,0,0
+user_521,Support,Analyst,Remote,SAP,31,1,1,0.481,0.67,30,15,1,11,1
+user_522,Support,Specialist,NY,Jira,19,1,0,0.287,0.937,92,11,10,4,0
+user_523,Finance,Analyst,NY,SAP,27,1,0,0.418,0.426,68,13,1,5,0
+user_524,Finance,Analyst,NY,Oracle,44,0,1,0.588,0.99,30,12,4,6,0
+user_525,Support,Manager,Remote,SAP,5,0,0,0.338,0.549,67,8,3,12,1
+user_526,Engineering,Admin,TX,GitHub,21,0,0,0.901,0.481,123,20,10,6,0
+user_527,Sales,Manager,TX,SAP,35,1,1,0.647,0.678,160,1,5,2,1
+user_528,Sales,Manager,NY,SAP,18,0,1,0.841,0.461,44,8,6,5,1
+user_529,HR,Engineer,NY,Workday,5,1,1,0.244,0.515,107,10,0,7,0
+user_530,HR,Admin,TX,GitHub,45,1,1,0.224,0.337,136,18,7,2,0
+user_531,HR,Admin,SF,Oracle,14,0,1,0.266,0.613,15,6,10,10,1
+user_532,Finance,Analyst,NY,Jira,31,0,1,0.74,0.76,161,13,7,11,1
+user_533,Engineering,Admin,SF,Workday,38,1,0,0.365,0.892,176,3,8,8,0
+user_534,HR,Admin,Remote,Workday,6,1,1,0.401,0.449,97,13,0,11,0
+user_535,Engineering,Engineer,NY,Oracle,10,0,1,0.75,0.303,106,9,2,12,1
+user_536,Sales,Manager,SF,Jira,31,0,1,0.831,0.378,99,13,1,12,0
+user_537,Engineering,Admin,NY,SAP,32,0,0,0.378,0.926,129,1,6,12,0
+user_538,HR,Engineer,Remote,Oracle,38,1,1,0.547,0.85,21,1,10,3,0
+user_539,Finance,Engineer,SF,Oracle,43,1,1,0.761,0.543,100,18,8,7,1
+user_540,Finance,Specialist,NY,Workday,30,0,0,0.629,0.541,164,5,7,8,1
+user_541,HR,Manager,NY,SAP,17,1,0,0.617,0.304,11,13,10,4,1
+user_542,Engineering,Manager,Remote,Jira,19,1,0,0.34,0.807,26,2,9,0,0
+user_543,Engineering,Manager,SF,SAP,22,0,1,0.695,0.545,123,2,0,1,1
+user_544,Finance,Engineer,TX,Workday,10,0,1,0.74,0.303,92,3,8,10,1
+user_545,Finance,Engineer,Remote,Jira,8,0,0,0.814,0.802,72,3,8,1,1
+user_546,Finance,Manager,TX,Jira,34,1,1,0.241,0.799,17,5,0,7,0
+user_547,Engineering,Engineer,TX,Jira,9,1,1,0.789,0.917,98,19,3,8,1
+user_548,Sales,Engineer,SF,Jira,34,1,1,0.22,0.53,154,17,1,3,1
+user_549,Engineering,Admin,SF,Oracle,10,1,1,0.568,0.923,100,7,9,6,1
+user_550,HR,Engineer,NY,Oracle,41,0,1,0.76,0.726,165,9,0,1,1
+user_551,Sales,Engineer,Remote,Workday,45,1,1,0.309,0.826,83,15,6,9,0
+user_552,HR,Engineer,TX,GitHub,36,0,0,0.38,0.467,7,18,8,9,1
+user_553,Finance,Engineer,TX,Oracle,7,0,0,0.286,0.505,60,8,1,9,1
+user_554,Support,Analyst,NY,Jira,19,1,1,0.225,0.829,95,13,10,3,0
+user_555,Engineering,Specialist,NY,GitHub,30,0,1,0.482,0.676,106,19,5,6,1
+user_556,Finance,Specialist,NY,GitHub,7,1,0,0.53,0.434,143,11,4,3,0
+user_557,Support,Manager,SF,Workday,15,1,0,0.245,0.424,137,20,5,12,0
+user_558,Finance,Analyst,TX,GitHub,35,0,0,0.672,0.624,161,6,5,12,0
+user_559,Support,Engineer,TX,Oracle,24,0,1,0.208,0.334,13,5,8,9,1
+user_560,Finance,Analyst,NY,Workday,16,0,0,0.558,0.45,149,20,1,12,1
+user_561,Finance,Admin,TX,GitHub,42,0,1,0.81,0.777,35,8,1,4,0
+user_562,Engineering,Analyst,Remote,GitHub,12,1,1,0.466,0.422,5,2,9,2,1
+user_563,Support,Analyst,NY,Jira,28,1,0,0.647,0.331,162,20,8,6,0
+user_564,Sales,Specialist,TX,Workday,17,1,0,0.688,0.243,96,5,1,10,0
+user_565,Engineering,Specialist,TX,Oracle,45,0,0,0.375,0.41,81,5,5,0,1
+user_566,Support,Engineer,TX,SAP,33,1,1,0.669,0.265,31,14,10,0,1
+user_567,Engineering,Analyst,TX,Jira,45,0,1,0.404,0.402,101,9,3,2,1
+user_568,Finance,Engineer,SF,SAP,27,1,1,0.364,0.488,56,11,0,1,1
+user_569,HR,Analyst,TX,Oracle,23,1,1,0.249,0.779,178,6,1,5,0
+user_570,Engineering,Analyst,TX,Workday,37,0,0,0.524,0.543,133,6,2,5,0
+user_571,Engineering,Manager,SF,Oracle,9,1,1,0.694,0.879,106,16,2,9,0
+user_572,Finance,Admin,TX,Oracle,30,0,1,0.587,0.988,51,3,6,4,0
+user_573,Finance,Analyst,SF,Workday,9,0,1,0.732,0.524,124,18,0,12,1
+user_574,Support,Analyst,TX,GitHub,7,1,1,0.427,0.803,63,15,3,3,1
+user_575,HR,Engineer,SF,GitHub,33,0,1,0.604,0.306,165,19,8,7,1
+user_576,Finance,Engineer,TX,Oracle,24,1,1,0.472,0.572,45,17,1,7,1
+user_577,HR,Specialist,Remote,Jira,41,1,1,0.9,0.921,158,16,2,5,1
+user_578,HR,Engineer,TX,SAP,22,0,0,0.379,0.899,124,13,9,3,1
+user_579,Engineering,Engineer,TX,Jira,26,1,0,0.612,0.967,118,14,7,2,0
+user_580,Sales,Analyst,TX,Jira,11,0,0,0.481,0.328,10,16,7,3,0
+user_581,Support,Analyst,Remote,SAP,44,0,1,0.229,0.409,14,8,8,1,1
+user_582,Finance,Engineer,TX,Oracle,33,1,0,0.409,0.63,119,7,6,8,1
+user_583,Engineering,Specialist,Remote,Jira,13,0,1,0.937,0.417,56,1,9,11,1
+user_584,Sales,Analyst,SF,Workday,41,1,0,0.552,0.34,2,15,10,11,1
+user_585,Finance,Manager,SF,SAP,18,1,0,0.552,0.511,67,9,6,11,0
+user_586,Support,Analyst,NY,Oracle,34,1,0,0.834,0.534,176,6,0,7,1
+user_587,Finance,Analyst,SF,SAP,22,1,1,0.607,0.605,18,13,9,0,1
+user_588,Engineering,Analyst,Remote,SAP,33,1,0,0.715,0.181,117,19,6,1,1
+user_589,Sales,Manager,TX,GitHub,23,0,0,0.82,0.239,80,17,9,7,0
+user_590,Finance,Admin,NY,GitHub,33,0,0,0.22,0.698,24,5,0,10,0
+user_591,HR,Analyst,Remote,GitHub,5,0,0,0.578,0.506,150,17,6,2,0
+user_592,HR,Admin,TX,Oracle,10,0,0,0.909,0.132,52,18,0,10,0
+user_593,Support,Manager,SF,SAP,39,1,0,0.451,0.255,157,12,8,11,1
+user_594,Sales,Specialist,TX,Oracle,29,0,1,0.781,0.585,15,2,7,6,1
+user_595,Sales,Admin,Remote,SAP,18,1,0,0.7,0.869,119,17,2,6,0
+user_596,HR,Specialist,TX,GitHub,17,0,0,0.567,0.74,59,16,9,1,1
+user_597,Sales,Engineer,SF,GitHub,34,0,0,0.91,0.739,105,20,0,1,0
+user_598,Sales,Manager,SF,Workday,35,1,1,0.799,0.88,113,15,8,9,0
+user_599,Finance,Manager,NY,Jira,36,1,1,0.218,0.258,91,14,4,10,1
+user_600,Engineering,Analyst,SF,Oracle,22,0,1,0.684,0.482,115,19,9,4,0
+user_601,Engineering,Engineer,TX,SAP,23,1,1,0.321,0.273,134,16,6,12,0
+user_602,Sales,Analyst,Remote,Oracle,29,1,1,0.654,0.896,118,14,0,1,1
+user_603,Finance,Engineer,TX,GitHub,7,0,1,0.54,0.562,102,2,8,0,0
+user_604,Finance,Manager,Remote,Jira,6,1,1,0.364,0.624,28,17,7,1,1
+user_605,Support,Analyst,NY,Jira,18,0,1,0.943,0.526,0,19,1,9,1
+user_606,Finance,Specialist,SF,SAP,42,0,1,0.335,0.475,126,8,8,4,1
+user_607,Engineering,Specialist,NY,Workday,42,1,1,0.829,0.203,26,6,5,6,0
+user_608,Finance,Engineer,TX,Oracle,32,1,0,0.799,0.841,169,6,3,4,0
+user_609,Support,Specialist,SF,Oracle,29,1,0,0.512,0.302,160,14,10,7,1
+user_610,HR,Analyst,SF,Oracle,43,1,1,0.819,0.712,85,0,6,9,0
+user_611,HR,Analyst,TX,Jira,26,1,1,0.334,0.154,80,4,7,5,1
+user_612,HR,Specialist,TX,GitHub,27,1,1,0.212,0.961,82,19,8,11,1
+user_613,Finance,Specialist,TX,Oracle,11,0,1,0.284,0.111,98,3,4,12,1
+user_614,HR,Analyst,NY,Jira,15,1,0,0.259,0.28,55,0,7,2,1
+user_615,Support,Admin,Remote,SAP,21,1,0,0.65,0.78,146,7,3,0,1
+user_616,Support,Specialist,NY,SAP,24,1,1,0.441,0.198,180,19,10,5,0
+user_617,Support,Analyst,SF,SAP,13,1,0,0.599,0.73,66,2,0,7,1
+user_618,Support,Admin,TX,GitHub,31,0,0,0.869,0.742,154,11,4,3,0
+user_619,Support,Admin,SF,GitHub,6,1,0,0.203,0.686,166,14,4,5,1
+user_620,Sales,Specialist,TX,Workday,33,1,1,0.205,0.406,29,3,9,9,1
+user_621,Engineering,Engineer,Remote,Oracle,31,1,1,0.454,0.312,36,17,9,6,1
+user_622,Support,Analyst,TX,Jira,42,1,1,0.793,0.982,122,13,2,4,1
+user_623,HR,Manager,TX,Workday,16,0,0,0.945,0.469,169,7,0,3,1
+user_624,Sales,Engineer,NY,Oracle,8,1,0,0.508,0.578,172,10,1,5,0
+user_625,Engineering,Engineer,SF,Oracle,35,1,0,0.605,0.591,112,4,1,1,0
+user_626,Sales,Manager,Remote,SAP,30,1,1,0.277,0.503,134,16,0,6,1
+user_627,Finance,Manager,NY,GitHub,5,1,1,0.912,0.264,160,14,7,6,0
+user_628,HR,Specialist,TX,GitHub,44,1,1,0.64,0.923,127,7,3,3,1
+user_629,Sales,Manager,SF,Oracle,13,1,0,0.53,0.485,40,8,2,8,0
+user_630,HR,Admin,Remote,Workday,18,1,0,0.577,0.414,41,18,1,1,0
+user_631,Sales,Engineer,NY,Oracle,9,1,1,0.873,0.725,63,15,10,8,1
+user_632,Finance,Specialist,TX,Jira,25,1,0,0.532,0.3,80,0,1,6,0
+user_633,Support,Analyst,SF,Oracle,38,0,1,0.662,0.771,12,10,0,5,1
+user_634,Sales,Admin,TX,SAP,38,1,1,0.727,0.223,151,5,0,6,0
+user_635,Finance,Specialist,Remote,GitHub,14,1,1,0.537,0.646,103,3,5,9,0
+user_636,HR,Admin,TX,GitHub,32,1,1,0.522,0.566,7,5,7,11,1
+user_637,HR,Specialist,TX,Jira,24,0,0,0.803,0.885,0,3,4,0,1
+user_638,Sales,Engineer,TX,GitHub,28,0,0,0.217,0.35,45,19,3,0,0
+user_639,Sales,Manager,Remote,GitHub,35,1,0,0.925,0.304,135,1,10,5,0
+user_640,HR,Engineer,SF,GitHub,13,1,0,0.547,0.514,179,6,4,11,1
+user_641,Finance,Manager,Remote,Workday,45,0,1,0.388,0.502,33,14,7,3,1
+user_642,Finance,Specialist,Remote,Workday,19,0,1,0.281,0.159,38,11,7,1,1
+user_643,Finance,Engineer,TX,Oracle,5,0,1,0.245,0.438,36,20,6,12,1
+user_644,HR,Specialist,SF,Jira,28,0,0,0.662,0.546,149,15,10,10,1
+user_645,Sales,Manager,Remote,Oracle,21,1,1,0.467,0.915,18,10,8,7,1
+user_646,HR,Specialist,NY,Workday,24,1,0,0.699,0.539,106,7,7,10,0
+user_647,Engineering,Admin,NY,Workday,29,1,0,0.785,0.165,131,15,3,6,0
+user_648,Sales,Admin,SF,Jira,13,1,0,0.334,0.643,47,12,10,7,0
+user_649,HR,Admin,Remote,SAP,37,1,0,0.342,0.558,120,11,0,3,0
+user_650,Finance,Analyst,NY,Jira,19,1,1,0.702,0.237,47,1,6,5,1
+user_651,Sales,Manager,TX,Oracle,11,1,1,0.498,0.277,52,4,3,11,0
+user_652,Sales,Analyst,NY,GitHub,36,1,1,0.316,0.143,122,16,2,3,1
+user_653,Support,Specialist,Remote,Jira,31,0,1,0.935,0.276,32,4,7,9,1
+user_654,Finance,Admin,NY,Oracle,41,1,1,0.664,0.348,127,19,0,0,1
+user_655,Finance,Specialist,NY,Oracle,33,0,1,0.682,0.453,143,7,5,6,0
+user_656,Sales,Analyst,SF,Jira,44,1,0,0.601,0.602,1,8,7,8,0
+user_657,Engineering,Engineer,TX,Jira,26,1,1,0.256,0.59,171,12,0,5,0
+user_658,Sales,Specialist,Remote,SAP,23,1,1,0.937,0.119,179,9,3,2,0
+user_659,Finance,Analyst,Remote,SAP,8,1,0,0.671,0.591,44,18,10,10,1
+user_660,Engineering,Admin,TX,Jira,17,0,0,0.927,0.738,101,0,0,5,1
+user_661,Finance,Admin,Remote,SAP,29,0,1,0.913,0.571,33,4,3,8,0
+user_662,Sales,Admin,TX,Oracle,9,1,0,0.937,0.689,105,1,5,6,1
+user_663,Engineering,Specialist,TX,Workday,12,1,1,0.906,0.137,75,11,3,0,0
+user_664,Sales,Analyst,TX,GitHub,36,0,1,0.547,0.903,156,14,5,11,0
+user_665,Support,Analyst,Remote,SAP,38,1,0,0.468,0.787,165,3,10,7,0
+user_666,HR,Admin,NY,Oracle,14,1,1,0.404,0.435,118,8,8,9,1
+user_667,Support,Specialist,SF,Oracle,22,1,1,0.487,0.659,30,5,2,12,0
+user_668,Sales,Admin,TX,SAP,18,0,1,0.546,0.222,178,1,7,4,1
+user_669,HR,Specialist,TX,SAP,22,0,1,0.283,0.891,146,4,1,10,1
+user_670,Engineering,Specialist,TX,Oracle,18,0,1,0.296,0.927,148,14,7,6,1
+user_671,Finance,Analyst,TX,SAP,21,1,0,0.435,0.754,6,12,2,10,0
+user_672,Sales,Specialist,SF,GitHub,20,1,0,0.77,0.145,163,5,9,1,0
+user_673,Sales,Admin,NY,GitHub,24,0,0,0.89,0.777,176,14,1,8,0
+user_674,Finance,Engineer,NY,Workday,44,1,1,0.884,0.162,45,20,6,8,1
+user_675,Support,Manager,SF,Jira,13,1,0,0.874,0.692,72,10,4,9,1
+user_676,Sales,Admin,Remote,SAP,25,0,0,0.221,0.807,59,4,10,11,0
+user_677,Sales,Engineer,SF,Jira,25,1,1,0.25,0.625,163,5,2,4,1
+user_678,Sales,Specialist,Remote,Oracle,30,1,1,0.252,0.383,77,16,1,3,1
+user_679,Support,Admin,SF,Oracle,12,1,1,0.875,0.592,20,9,0,1,0
+user_680,Finance,Engineer,SF,Workday,33,1,1,0.547,0.567,174,14,8,2,0
+user_681,Finance,Manager,SF,Workday,30,1,0,0.376,0.873,9,0,6,2,0
+user_682,HR,Admin,NY,GitHub,31,1,0,0.665,0.679,157,5,7,12,1
+user_683,Support,Engineer,Remote,GitHub,38,0,1,0.604,0.743,159,19,3,12,0
+user_684,Finance,Engineer,Remote,Jira,31,1,0,0.659,0.12,134,19,3,5,0
+user_685,Support,Specialist,Remote,Oracle,41,0,1,0.631,0.549,138,6,2,8,0
+user_686,Sales,Specialist,TX,GitHub,12,1,1,0.538,0.872,133,10,6,12,0
+user_687,Engineering,Manager,Remote,GitHub,7,1,1,0.552,0.873,169,16,1,0,0
+user_688,Finance,Engineer,Remote,Workday,38,0,1,0.615,0.352,69,19,4,10,1
+user_689,Support,Manager,TX,Workday,17,1,1,0.461,0.954,3,8,1,10,1
+user_690,HR,Engineer,SF,Workday,12,0,0,0.272,0.86,92,10,8,11,1
+user_691,Finance,Engineer,TX,GitHub,18,0,0,0.222,0.12,48,11,0,2,0
+user_692,HR,Analyst,Remote,Jira,37,0,0,0.533,0.621,6,19,4,6,0
+user_693,Support,Analyst,Remote,GitHub,34,1,0,0.544,0.512,50,14,9,8,0
+user_694,Support,Manager,NY,GitHub,41,0,1,0.538,0.519,81,15,5,7,1
+user_695,Engineering,Analyst,NY,Oracle,40,1,0,0.755,0.37,27,15,6,2,1
+user_696,Sales,Specialist,TX,Workday,13,0,1,0.768,0.68,169,4,1,1,1
+user_697,Finance,Analyst,SF,Oracle,21,0,1,0.505,0.237,105,13,9,8,1
+user_698,Support,Engineer,TX,GitHub,23,0,1,0.918,0.813,156,18,9,8,1
+user_699,Engineering,Manager,SF,Oracle,18,0,0,0.55,0.298,77,13,6,12,0
+user_700,Engineering,Specialist,SF,Workday,24,0,1,0.687,0.431,125,7,5,6,0
+user_701,Finance,Engineer,TX,SAP,12,0,0,0.655,0.64,145,12,0,10,1
+user_702,Engineering,Manager,TX,Jira,30,1,1,0.274,0.674,86,14,10,12,0
+user_703,Engineering,Engineer,NY,GitHub,30,0,0,0.794,0.734,37,13,6,9,1
+user_704,Support,Analyst,SF,Oracle,35,0,1,0.357,0.931,97,4,9,5,0
+user_705,Sales,Specialist,TX,Jira,15,1,0,0.338,0.94,141,13,5,11,1
+user_706,HR,Manager,SF,Oracle,39,0,1,0.693,0.73,142,19,4,11,1
+user_707,Engineering,Specialist,TX,Oracle,32,0,1,0.656,0.554,171,2,8,7,1
+user_708,Sales,Analyst,TX,Workday,15,0,0,0.619,0.242,110,20,6,0,1
+user_709,Support,Specialist,TX,GitHub,41,1,0,0.82,0.978,108,0,10,6,0
+user_710,Engineering,Manager,Remote,Oracle,29,1,0,0.227,0.648,158,6,2,9,0
+user_711,Engineering,Engineer,TX,GitHub,34,0,1,0.556,0.636,174,20,8,7,0
+user_712,Support,Specialist,Remote,GitHub,16,1,0,0.817,0.794,29,8,2,9,0
+user_713,Sales,Manager,SF,GitHub,14,1,0,0.939,0.965,135,8,9,3,0
+user_714,HR,Manager,NY,Jira,43,1,1,0.48,0.289,81,16,1,3,1
+user_715,Sales,Manager,SF,GitHub,28,0,1,0.762,0.799,10,13,8,2,1
+user_716,Finance,Engineer,TX,SAP,14,0,1,0.876,0.197,133,3,9,1,0
+user_717,HR,Specialist,SF,GitHub,31,1,1,0.61,0.235,57,19,8,7,1
+user_718,Sales,Engineer,SF,Jira,43,1,0,0.264,0.989,30,9,5,11,1
+user_719,Engineering,Admin,TX,Jira,10,0,0,0.433,0.792,136,7,6,11,0
+user_720,Finance,Engineer,NY,Workday,15,1,0,0.787,0.752,96,17,4,2,1
+user_721,Sales,Analyst,TX,GitHub,28,1,1,0.379,0.822,17,0,7,12,1
+user_722,Support,Analyst,SF,GitHub,44,0,1,0.431,0.684,104,9,6,8,1
+user_723,Support,Engineer,TX,SAP,23,1,0,0.461,0.651,53,12,1,9,1
+user_724,Engineering,Specialist,Remote,SAP,16,0,1,0.628,0.976,98,13,5,10,0
+user_725,Support,Engineer,Remote,GitHub,11,1,1,0.246,0.484,126,1,0,8,1
+user_726,Engineering,Admin,Remote,GitHub,24,0,1,0.563,0.363,175,13,1,1,0
+user_727,Finance,Engineer,NY,GitHub,18,0,0,0.658,0.354,166,13,7,0,1
+user_728,Engineering,Manager,Remote,GitHub,12,1,1,0.752,0.349,137,16,5,9,0
+user_729,Support,Admin,TX,GitHub,18,0,1,0.779,0.836,129,14,1,5,1
+user_730,Support,Manager,Remote,SAP,19,1,1,0.21,0.614,139,11,9,11,1
+user_731,Sales,Engineer,SF,GitHub,11,1,1,0.38,0.631,53,11,10,5,0
+user_732,Sales,Manager,SF,SAP,22,1,0,0.892,0.434,56,12,5,9,1
+user_733,Finance,Admin,SF,SAP,38,0,0,0.485,0.182,80,18,8,4,1
+user_734,HR,Specialist,Remote,SAP,29,0,1,0.903,0.667,98,11,1,6,1
+user_735,Sales,Manager,SF,Oracle,28,0,1,0.921,0.297,52,13,5,7,0
+user_736,HR,Engineer,Remote,Jira,41,1,0,0.202,0.135,154,1,6,5,0
+user_737,Engineering,Manager,NY,SAP,11,1,1,0.639,0.419,58,7,2,0,0
+user_738,Finance,Specialist,TX,Jira,6,1,1,0.582,0.87,11,7,7,9,1
+user_739,Sales,Analyst,SF,GitHub,21,0,1,0.547,0.295,55,18,10,5,1
+user_740,HR,Analyst,TX,Workday,17,1,0,0.747,0.635,165,4,10,7,0
+user_741,Engineering,Admin,Remote,Jira,18,1,1,0.383,0.686,174,7,10,8,1
+user_742,Finance,Engineer,Remote,Workday,17,0,1,0.538,0.669,141,3,6,11,1
+user_743,Sales,Engineer,NY,SAP,18,0,0,0.256,0.758,111,13,10,5,1
+user_744,Finance,Admin,NY,Workday,7,1,1,0.834,0.742,1,6,6,8,1
+user_745,HR,Specialist,TX,Oracle,10,0,1,0.472,0.369,160,19,0,0,0
+user_746,HR,Admin,Remote,Workday,14,1,1,0.947,0.91,10,11,4,8,1
+user_747,HR,Engineer,SF,SAP,9,0,0,0.311,0.511,51,20,5,11,0
+user_748,HR,Specialist,NY,Jira,32,1,1,0.596,0.401,174,6,3,5,0
+user_749,HR,Admin,SF,Workday,6,0,0,0.795,0.686,136,10,1,2,1
+user_750,Engineering,Specialist,Remote,Oracle,23,0,1,0.733,0.765,125,14,0,4,0
+user_751,Finance,Manager,Remote,SAP,24,0,1,0.549,0.142,46,14,5,5,1
+user_752,Sales,Analyst,Remote,Workday,13,1,0,0.723,0.768,102,4,7,5,0
+user_753,Engineering,Engineer,SF,GitHub,14,0,1,0.468,0.594,66,0,1,5,1
+user_754,Support,Manager,NY,Workday,35,1,1,0.497,0.944,122,11,10,12,0
+user_755,Sales,Manager,Remote,GitHub,42,1,0,0.507,0.116,1,2,3,5,1
+user_756,Finance,Analyst,NY,Workday,10,0,1,0.764,0.137,8,13,2,12,1
+user_757,Finance,Admin,SF,Workday,40,1,1,0.421,0.833,27,7,9,3,0
+user_758,Engineering,Manager,Remote,Oracle,27,1,0,0.908,0.953,61,6,0,10,1
+user_759,HR,Admin,NY,GitHub,36,1,0,0.773,0.7,67,11,0,4,1
+user_760,Engineering,Analyst,Remote,GitHub,10,0,0,0.617,0.659,97,10,1,2,0
+user_761,Finance,Engineer,SF,GitHub,11,1,0,0.779,0.523,86,16,1,8,0
+user_762,Support,Analyst,Remote,Oracle,37,0,0,0.682,0.661,139,5,8,6,1
+user_763,Sales,Admin,SF,GitHub,16,1,1,0.411,0.843,162,20,0,8,1
+user_764,Support,Admin,Remote,SAP,15,1,0,0.316,0.585,19,10,8,7,0
+user_765,Support,Analyst,SF,GitHub,20,0,0,0.659,0.385,147,18,5,12,0
+user_766,Support,Analyst,SF,Oracle,34,1,1,0.379,0.25,139,6,7,7,0
+user_767,Finance,Manager,TX,SAP,38,0,1,0.853,0.887,57,11,3,5,0
+user_768,HR,Analyst,SF,Workday,7,1,1,0.863,0.603,176,7,10,10,0
+user_769,Engineering,Manager,TX,Jira,36,1,0,0.619,0.201,71,15,6,3,0
+user_770,Sales,Analyst,TX,SAP,44,1,1,0.634,0.909,163,8,2,6,0
+user_771,Finance,Admin,Remote,Jira,29,0,0,0.505,0.484,106,4,2,7,1
+user_772,HR,Analyst,TX,Jira,28,1,1,0.611,0.104,60,14,9,8,1
+user_773,Engineering,Engineer,SF,GitHub,12,1,0,0.217,0.961,164,13,7,2,1
+user_774,Sales,Analyst,NY,Jira,32,1,1,0.918,0.351,178,1,7,2,0
+user_775,Sales,Admin,NY,Oracle,39,0,0,0.883,0.426,70,7,2,0,1
+user_776,Sales,Specialist,SF,SAP,5,1,0,0.714,0.243,15,10,8,0,0
+user_777,HR,Specialist,Remote,Oracle,5,0,1,0.535,0.948,137,13,8,1,1
+user_778,HR,Engineer,NY,Oracle,42,0,0,0.203,0.237,167,10,2,6,1
+user_779,Support,Analyst,TX,SAP,35,0,0,0.941,0.987,70,8,4,2,0
+user_780,Finance,Analyst,SF,Oracle,32,1,0,0.454,0.255,58,17,7,12,0
+user_781,HR,Analyst,SF,SAP,12,1,0,0.591,0.339,78,13,10,6,0
+user_782,Support,Admin,SF,Workday,45,1,1,0.656,0.898,65,10,3,5,1
+user_783,Engineering,Admin,NY,Oracle,30,1,1,0.903,0.934,157,4,5,2,1
+user_784,Finance,Analyst,SF,SAP,26,0,0,0.2,0.399,16,13,7,3,0
+user_785,HR,Admin,SF,Workday,27,0,1,0.315,0.574,172,17,4,6,1
+user_786,Sales,Analyst,Remote,GitHub,24,1,1,0.738,0.46,153,12,2,9,0
+user_787,Support,Manager,SF,Jira,11,0,1,0.348,0.58,66,9,2,0,1
+user_788,HR,Admin,Remote,SAP,30,0,1,0.707,0.258,96,9,8,2,1
+user_789,Support,Analyst,SF,Oracle,33,0,0,0.872,0.126,37,16,3,8,1
+user_790,HR,Analyst,TX,SAP,10,1,1,0.249,0.425,141,3,1,11,0
+user_791,Support,Engineer,TX,Jira,29,0,0,0.396,0.375,118,20,9,12,1
+user_792,Engineering,Specialist,Remote,GitHub,29,1,0,0.695,0.812,63,4,10,5,1
+user_793,Support,Specialist,NY,SAP,39,1,1,0.669,0.534,44,1,4,0,0
+user_794,Support,Admin,NY,Oracle,37,0,1,0.519,0.231,58,1,5,5,0
+user_795,Support,Specialist,SF,GitHub,11,0,1,0.509,0.283,161,2,1,10,0
+user_796,Finance,Analyst,Remote,Jira,41,0,0,0.706,0.175,8,9,0,11,1
+user_797,HR,Analyst,TX,GitHub,27,1,0,0.475,0.885,90,9,9,8,1
+user_798,HR,Analyst,Remote,Workday,11,1,0,0.824,0.316,43,3,5,4,1
+user_799,Finance,Engineer,TX,Oracle,33,1,1,0.576,0.967,103,5,10,5,0

requirements.txt

@@ -0,0 +1,9 @@
+fastapi==0.110.0
+uvicorn==0.29.0
+pandas==2.2.2
+scikit-learn==1.4.2
+numpy==1.26.4
+sqlalchemy==2.0.29
+python-multipart==0.0.9
+openpyxl==3.1.2
+hdbscan==0.8.40