churnkit 0.75.0a1__py3-none-any.whl

churnkit-0.75.0a1.data/data/share/churnkit/exploration_notebooks/01b_temporal_quality.ipynb

@@ -0,0 +1,679 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "cell-0",
+   "metadata": {
+    "papermill": {
+     "duration": 0.002487,
+     "end_time": "2026-02-02T13:00:52.603186",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:52.600699",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "# Chapter 1b: Temporal Quality Assessment (Event Bronze Track)\n",
+    "\n",
+    "**Purpose:** Run quality checks specific to event-level datasets to identify data issues before feature engineering.\n",
+    "\n",
+    "**When to use this notebook:**\n",
+    "- After completing 01a_temporal_deep_dive.ipynb\n",
+    "- Your dataset is EVENT_LEVEL granularity\n",
+    "- You want to validate temporal data integrity before aggregation\n",
+    "\n",
+    "| Check | What It Detects | Why It Matters for ML |\n",
+    "|-------|-----------------|----------------------|\n",
+    "| **TQ001** | Duplicate events (same entity + timestamp) | Inflates counts, skews aggregations, creates artificial sequence patterns |\n",
+    "| **TQ002** | Unexpected temporal gaps | Rolling features become misleading; \"events in last 30d\" drops during gaps |\n",
+    "| **TQ003** | Future dates | Data leakage — model sees future during training |\n",
+    "| **TQ004** | Ambiguous event ordering | Sequence features undefined when multiple events share timestamp |"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-1",
+   "metadata": {
+    "papermill": {
+     "duration": 0.00158,
+     "end_time": "2026-02-02T13:00:52.606755",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:52.605175",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.1 Load Findings and Data"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-2",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:52.611011Z",
+     "iopub.status.busy": "2026-02-02T13:00:52.610876Z",
+     "iopub.status.idle": "2026-02-02T13:00:54.504186Z",
+     "shell.execute_reply": "2026-02-02T13:00:54.503492Z"
+    },
+    "papermill": {
+     "duration": 1.896686,
+     "end_time": "2026-02-02T13:00:54.505137",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:52.608451",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "from customer_retention.analysis.notebook_progress import track_and_export_previous\n",
+    "track_and_export_previous(\"01b_temporal_quality.ipynb\")\n",
+    "\n",
+    "from pathlib import Path\n",
+    "import pandas as pd\n",
+    "import plotly.graph_objects as go\n",
+    "from plotly.subplots import make_subplots\n",
+    "\n",
+    "from customer_retention.analysis.auto_explorer import ExplorationFindings, RecommendationEngine\n",
+    "from customer_retention.analysis.visualization import ChartBuilder, display_figure\n",
+    "from customer_retention.core.config.column_config import ColumnType\n",
+    "from customer_retention.stages.profiling import (\n",
+    "    DuplicateEventCheck, TemporalGapCheck, FutureDateCheck, EventOrderCheck,\n",
+    "    TemporalQualityReporter, SegmentAwareOutlierAnalyzer\n",
+    ")\n",
+    "from customer_retention.stages.temporal import load_data_with_snapshot_preference, TEMPORAL_METADATA_COLS\n",
+    "from customer_retention.core.config.experiments import FINDINGS_DIR, EXPERIMENTS_DIR, OUTPUT_DIR, setup_experiments_structure\n"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-3",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:54.509631Z",
+     "iopub.status.busy": "2026-02-02T13:00:54.509304Z",
+     "iopub.status.idle": "2026-02-02T13:00:55.198236Z",
+     "shell.execute_reply": "2026-02-02T13:00:55.197651Z"
+    },
+    "papermill": {
+     "duration": 0.692154,
+     "end_time": "2026-02-02T13:00:55.199216",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:54.507062",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# FINDINGS_DIR imported from customer_retention.core.config.experiments\n",
+    "findings_files = sorted(\n",
+    "    [f for f in FINDINGS_DIR.glob(\"*_findings.yaml\") if \"multi_dataset\" not in f.name],\n",
+    "    key=lambda f: f.stat().st_mtime, reverse=True\n",
+    ")\n",
+    "if not findings_files:\n",
+    "    raise FileNotFoundError(f\"No findings in {FINDINGS_DIR}. Run notebook 01 first.\")\n",
+    "\n",
+    "FINDINGS_PATH = str(findings_files[0])\n",
+    "findings = ExplorationFindings.load(FINDINGS_PATH)\n",
+    "print(f\"Using: {FINDINGS_PATH}\")\n",
+    "\n",
+    "ts_meta = findings.time_series_metadata\n",
+    "ENTITY_COLUMN, TIME_COLUMN = ts_meta.entity_column, ts_meta.time_column\n",
+    "print(f\"Entity: {ENTITY_COLUMN}, Time: {TIME_COLUMN}\")\n",
+    "\n",
+    "df, data_source = load_data_with_snapshot_preference(findings, output_dir=str(FINDINGS_DIR))\n",
+    "charts = ChartBuilder()\n",
+    "print(f\"Loaded {len(df):,} rows ({data_source})\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-4",
+   "metadata": {
+    "papermill": {
+     "duration": 0.0014,
+     "end_time": "2026-02-02T13:00:55.202447",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.201047",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.2 Configure Quality Checks"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-5",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:55.206889Z",
+     "iopub.status.busy": "2026-02-02T13:00:55.206640Z",
+     "iopub.status.idle": "2026-02-02T13:00:55.209364Z",
+     "shell.execute_reply": "2026-02-02T13:00:55.208747Z"
+    },
+    "papermill": {
+     "duration": 0.005584,
+     "end_time": "2026-02-02T13:00:55.209814",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.204230",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "REFERENCE_DATE = pd.Timestamp.now()  # or pd.Timestamp(\"2024-01-01\")\n",
+    "EXPECTED_FREQUENCY = \"D\"  # D=daily, W=weekly, M=monthly, H=hourly\n",
+    "MAX_GAP_MULTIPLE = 3.0\n",
+    "\n",
+    "print(f\"Reference: {REFERENCE_DATE.date()}, Frequency: {EXPECTED_FREQUENCY}, Gap threshold: {MAX_GAP_MULTIPLE}x\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-6",
+   "metadata": {
+    "papermill": {
+     "duration": 0.001492,
+     "end_time": "2026-02-02T13:00:55.213170",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.211678",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.3 Run Temporal Quality Checks\n",
+    "\n",
+    "| Issue Type | ML Impact | Mitigation |\n",
+    "|------------|-----------|------------|\n",
+    "| Duplicates | Sum/count features inflated; artificial patterns in sequences | Deduplicate or add sequence index |\n",
+    "| Gaps | Rolling aggregations drop; recency features spike | Document gaps; add gap indicator feature |\n",
+    "| Future dates | Model trains on leaked future info | Filter to reference date; check timezone handling |\n",
+    "| Ordering | \"Previous event\" features undefined | Add tiebreaker column; use stable sort |"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-7",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:55.217026Z",
+     "iopub.status.busy": "2026-02-02T13:00:55.216922Z",
+     "iopub.status.idle": "2026-02-02T13:00:55.267263Z",
+     "shell.execute_reply": "2026-02-02T13:00:55.266748Z"
+    },
+    "papermill": {
+     "duration": 0.053177,
+     "end_time": "2026-02-02T13:00:55.267824",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.214647",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "checks = [\n",
+    "    DuplicateEventCheck(entity_column=ENTITY_COLUMN, time_column=TIME_COLUMN),\n",
+    "    TemporalGapCheck(time_column=TIME_COLUMN, expected_frequency=EXPECTED_FREQUENCY, max_gap_multiple=MAX_GAP_MULTIPLE),\n",
+    "    FutureDateCheck(time_column=TIME_COLUMN, reference_date=REFERENCE_DATE),\n",
+    "    EventOrderCheck(entity_column=ENTITY_COLUMN, time_column=TIME_COLUMN),\n",
+    "]\n",
+    "results = [check.run(df) for check in checks]\n",
+    "reporter = TemporalQualityReporter(results, len(df))\n",
+    "reporter.print_results()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-8",
+   "metadata": {
+    "papermill": {
+     "duration": 0.001479,
+     "end_time": "2026-02-02T13:00:55.270977",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.269498",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.4 Quality Score\n",
+    "\n",
+    "| Component | Weight | Scoring Logic |\n",
+    "|-----------|--------|---------------|\n",
+    "| Each check | 25% | 100 if no issues; deductions proportional to % affected |\n",
+    "| Grade A | 90-100 | Proceed with confidence |\n",
+    "| Grade B | 75-89 | Document issues, proceed with caution |\n",
+    "| Grade C | 60-74 | Address issues before feature engineering |\n",
+    "| Grade D | <60 | Investigation required |"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-9",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:55.275294Z",
+     "iopub.status.busy": "2026-02-02T13:00:55.275192Z",
+     "iopub.status.idle": "2026-02-02T13:00:55.277551Z",
+     "shell.execute_reply": "2026-02-02T13:00:55.276851Z"
+    },
+    "papermill": {
+     "duration": 0.005227,
+     "end_time": "2026-02-02T13:00:55.278101",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.272874",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "reporter.print_score()\n",
+    "quality_score, grade, passed = reporter.quality_score, reporter.grade, reporter.passed"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-10",
+   "metadata": {
+    "papermill": {
+     "duration": 0.001615,
+     "end_time": "2026-02-02T13:00:55.281714",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.280099",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.5 Event Volume Analysis\n",
+    "\n",
+    "| What to Look For | Indicates | Action |\n",
+    "|-----------------|-----------|--------|\n",
+    "| Missing bars | Data gaps (TQ002) | Document; add gap indicator |\n",
+    "| Declining trend | Population shrinkage or data cutoff | Check if intentional |\n",
+    "| Spikes | Campaigns, seasonality, or data issues | Investigate cause |\n",
+    "| Flat periods | Possible logging outages | Verify with data source |"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-11",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:55.286039Z",
+     "iopub.status.busy": "2026-02-02T13:00:55.285927Z",
+     "iopub.status.idle": "2026-02-02T13:00:55.315871Z",
+     "shell.execute_reply": "2026-02-02T13:00:55.315137Z"
+    },
+    "papermill": {
+     "duration": 0.033261,
+     "end_time": "2026-02-02T13:00:55.316776",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.283515",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "df_temp = df.copy()\n",
+    "df_temp[TIME_COLUMN] = pd.to_datetime(df_temp[TIME_COLUMN])\n",
+    "time_span = (df_temp[TIME_COLUMN].max() - df_temp[TIME_COLUMN].min()).days\n",
+    "\n",
+    "freq, label = (\"D\", \"Daily\") if time_span <= 90 else (\"W\", \"Weekly\") if time_span <= 365 else (\"ME\", \"Monthly\")\n",
+    "counts = df_temp.groupby(pd.Grouper(key=TIME_COLUMN, freq=freq)).size()\n",
+    "\n",
+    "fig = go.Figure(go.Bar(x=counts.index, y=counts.values, marker_color=\"#4682B4\"))\n",
+    "fig.update_layout(title=f\"{label} Event Volume (gaps = missing bars)\", height=300, template=\"plotly_white\")\n",
+    "display_figure(fig)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-14",
+   "metadata": {
+    "papermill": {
+     "duration": 0.003275,
+     "end_time": "2026-02-02T13:00:55.323645",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.320370",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.6 Outlier Analysis\n",
+    "\n",
+    "| Approach | When to Use | Why It Matters |\n",
+    "|----------|-------------|----------------|\n",
+    "| Global detection | Homogeneous data | Simple threshold works |\n",
+    "| Segment-aware | Data has natural groups | Avoids false positives when segments have different scales |\n",
+    "\n",
+    "Segment-aware detection clusters entities by target (or other segment) and detects outliers within each group separately."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-15",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:00:55.331243Z",
+     "iopub.status.busy": "2026-02-02T13:00:55.331125Z",
+     "iopub.status.idle": "2026-02-02T13:01:05.464597Z",
+     "shell.execute_reply": "2026-02-02T13:01:05.463713Z"
+    },
+    "papermill": {
+     "duration": 10.138458,
+     "end_time": "2026-02-02T13:01:05.465385",
+     "exception": false,
+     "start_time": "2026-02-02T13:00:55.326927",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "numeric_cols = [n for n, c in findings.columns.items()\n",
+    "    if c.inferred_type in [ColumnType.NUMERIC_CONTINUOUS, ColumnType.NUMERIC_DISCRETE]\n",
+    "    and n not in [ENTITY_COLUMN, TIME_COLUMN] and n not in TEMPORAL_METADATA_COLS]\n",
+    "\n",
+    "if numeric_cols:\n",
+    "    analyzer = SegmentAwareOutlierAnalyzer(max_segments=5)\n",
+    "    result = analyzer.analyze(df, feature_cols=numeric_cols, segment_col=None, target_col=findings.target_column)\n",
+    "    \n",
+    "    print(f\"Segments detected: {result.n_segments}\")\n",
+    "    if result.n_segments > 1:\n",
+    "        data = [{\"Feature\": c, \"Global\": result.global_analysis[c].outliers_detected,\n",
+    "            \"Segment\": sum(s[c].outliers_detected for s in result.segment_analysis.values() if c in s)}\n",
+    "            for c in numeric_cols]\n",
+    "        display(pd.DataFrame(data))\n",
+    "        if result.segmentation_recommended:\n",
+    "            print(\"\\n💡 Segment-specific outlier treatment recommended\")\n",
+    "    else:\n",
+    "        print(\"Data appears homogeneous - using global outlier detection\")\n",
+    "else:\n",
+    "    print(\"No numeric columns for outlier analysis.\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-16",
+   "metadata": {
+    "papermill": {
+     "duration": 0.003472,
+     "end_time": "2026-02-02T13:01:05.472843",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.469371",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.7 Data Validation\n",
+    "\n",
+    "| Check | Issue | Impact |\n",
+    "|-------|-------|--------|\n",
+    "| Binary fields | Values outside {0, 1} | Model crashes or silent errors |\n",
+    "| String consistency | Case/spacing variants (\"Yes\" vs \"yes\") | Inflated cardinality; split categories |\n",
+    "| Missing patterns | Systematic missingness | Bias in imputation |"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-17",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:01:05.481150Z",
+     "iopub.status.busy": "2026-02-02T13:01:05.481017Z",
+     "iopub.status.idle": "2026-02-02T13:01:05.648331Z",
+     "shell.execute_reply": "2026-02-02T13:01:05.647595Z"
+    },
+    "papermill": {
+     "duration": 0.172374,
+     "end_time": "2026-02-02T13:01:05.649066",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.476692",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "# Binary field validation\n",
+    "binary_cols = [n for n, c in findings.columns.items() if c.inferred_type == ColumnType.BINARY and n not in TEMPORAL_METADATA_COLS]\n",
+    "for col in binary_cols:\n",
+    "    c0, c1 = (df[col] == 0).sum(), (df[col] == 1).sum()\n",
+    "    print(f\"✓ {col}: 0={c0:,} ({c0/(c0+c1)*100:.1f}%), 1={c1:,} ({c1/(c0+c1)*100:.1f}%)\")\n",
+    "\n",
+    "# Consistency check\n",
+    "issues = []\n",
+    "for col in df.select_dtypes(include=['object']).columns:\n",
+    "    if col in [ENTITY_COLUMN, TIME_COLUMN]: continue\n",
+    "    variants = {}\n",
+    "    for v in df[col].dropna().unique():\n",
+    "        key = str(v).lower().strip()\n",
+    "        variants.setdefault(key, []).append(v)\n",
+    "    issues.extend([{\"Column\": col, \"Variants\": vs} for vs in variants.values() if len(vs) > 1])\n",
+    "\n",
+    "print(f\"\\n{'⚠️ Consistency issues: ' + str(len(issues)) if issues else '✅ No consistency issues'}\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-18",
+   "metadata": {
+    "papermill": {
+     "duration": 0.003229,
+     "end_time": "2026-02-02T13:01:05.655939",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.652710",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.8 Recommendations\n",
+    "\n",
+    "Framework-generated recommendations based on column-level issues detected during exploration."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-19",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:01:05.663919Z",
+     "iopub.status.busy": "2026-02-02T13:01:05.663801Z",
+     "iopub.status.idle": "2026-02-02T13:01:05.667016Z",
+     "shell.execute_reply": "2026-02-02T13:01:05.666376Z"
+    },
+    "papermill": {
+     "duration": 0.00802,
+     "end_time": "2026-02-02T13:01:05.667651",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.659631",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "rec_engine = RecommendationEngine()\n",
+    "recs = rec_engine.recommend_cleaning(findings)\n",
+    "\n",
+    "if recs:\n",
+    "    for r in sorted(recs, key=lambda x: {\"high\": 0, \"medium\": 1, \"low\": 2}.get(x.severity, 3)):\n",
+    "        icon = {\"high\": \"🔴\", \"medium\": \"🟡\", \"low\": \"🟢\"}.get(r.severity, \"⚪\")\n",
+    "        print(f\"{icon} [{r.severity.upper()}] {r.column_name}: {r.description}\")\n",
+    "        label = r.strategy_label if r.strategy_label else r.strategy.replace(\"_\", \" \").title()\n",
+    "        print(f\"   Strategy: {label}\")\n",
+    "else:\n",
+    "    print(\"✅ No critical cleaning recommendations\")"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-20",
+   "metadata": {
+    "papermill": {
+     "duration": 0.004024,
+     "end_time": "2026-02-02T13:01:05.675373",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.671349",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "## 1b.9 Save Results"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "cell-21",
+   "metadata": {
+    "execution": {
+     "iopub.execute_input": "2026-02-02T13:01:05.683302Z",
+     "iopub.status.busy": "2026-02-02T13:01:05.683171Z",
+     "iopub.status.idle": "2026-02-02T13:01:05.699141Z",
+     "shell.execute_reply": "2026-02-02T13:01:05.698285Z"
+    },
+    "papermill": {
+     "duration": 0.021202,
+     "end_time": "2026-02-02T13:01:05.699979",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.678777",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "outputs": [],
+   "source": [
+    "if not findings.metadata:\n",
+    "    findings.metadata = {}\n",
+    "findings.metadata[\"temporal_quality\"] = reporter.to_dict()\n",
+    "findings.save(FINDINGS_PATH)\n",
+    "print(f\"Saved to: {FINDINGS_PATH}\")\n",
+    "print(f\"Score: {quality_score:.0f}/100 (Grade {grade})\")\n"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "cell-22",
+   "metadata": {
+    "papermill": {
+     "duration": 0.003113,
+     "end_time": "2026-02-02T13:01:05.706587",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.703474",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "---\n",
+    "\n",
+    "## Summary: What We Learned\n",
+    "\n",
+    "In this notebook, we validated temporal data quality:\n",
+    "\n",
+    "1. **Temporal Quality Checks** — Detected duplicates, gaps, future dates, ordering issues\n",
+    "2. **Quality Score** — Quantified overall data health with pass/fail grading\n",
+    "3. **Event Volume** — Visualized data coverage over time\n",
+    "4. **Outlier Analysis** — Compared global vs segment-aware detection\n",
+    "5. **Data Validation** — Verified binary fields and string consistency\n",
+    "\n",
+    "## Quality Score Interpretation\n",
+    "\n",
+    "| Grade | Score | Meaning | Action |\n",
+    "|-------|-------|---------|--------|\n",
+    "| A | 90-100 | Excellent | Proceed with confidence |\n",
+    "| B | 75-89 | Good | Document issues, proceed |\n",
+    "| C | 60-74 | Fair | Address issues before aggregation |\n",
+    "| D | <60 | Poor | Investigation required |\n",
+    "\n",
+    "---\n",
+    "\n",
+    "## Next Steps\n",
+    "\n",
+    "Continue with the **Event Bronze Track**:\n",
+    "\n",
+    "1. **01c_temporal_patterns.ipynb** — Detect trends, seasonality, cohort effects\n",
+    "2. **01d_event_aggregation.ipynb** — Aggregate events to entity-level features\n",
+    "\n",
+    "After 01d, continue with **Entity Bronze Track** (02 → 03 → 04) on aggregated data."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "id": "f1ec2c4e",
+   "metadata": {
+    "papermill": {
+     "duration": 0.00297,
+     "end_time": "2026-02-02T13:01:05.712602",
+     "exception": false,
+     "start_time": "2026-02-02T13:01:05.709632",
+     "status": "completed"
+    },
+    "tags": []
+   },
+   "source": [
+    "> **Save Reminder:** Save this notebook (Ctrl+S / Cmd+S) before running the next one.\n",
+    "> The next notebook will automatically export this notebook's HTML documentation from the saved file."
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "language": "python",
+   "name": "python3"
+  },
+  "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
+   "name": "python",
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.12.4"
+  },
+  "papermill": {
+   "default_parameters": {},
+   "duration": 16.265903,
+   "end_time": "2026-02-02T13:01:08.333031",
+   "environment_variables": {},
+   "exception": null,
+   "input_path": "/Users/Vital/python/CustomerRetention/exploration_notebooks/01b_temporal_quality.ipynb",
+   "output_path": "/Users/Vital/python/CustomerRetention/exploration_notebooks/01b_temporal_quality.ipynb",
+   "parameters": {},
+   "start_time": "2026-02-02T13:00:52.067128",
+   "version": "2.6.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 5
+}