CWICR Historical Cost Tracker

Business Case

Problem Statement

Improving estimates requires:

• Actual cost feedback
• Historical comparisons
• Trend analysis
• Lessons learned

Solution

Track actual costs against CWICR estimates, build historical database, and use data to improve future estimating accuracy.

Business Value

• Accuracy improvement - Learn from actuals
• Benchmarking - Project comparisons
• Trend analysis - Cost movement patterns
• Organizational knowledge - Cost database

Technical Implementation

import pandas as pd
import numpy as np
from typing import Dict, Any, List, Optional
from dataclasses import dataclass, field
from datetime import datetime, date
from enum import Enum
import json


class ProjectStatus(Enum):
    """Project status."""
    ESTIMATED = "estimated"
    IN_PROGRESS = "in_progress"
    COMPLETED = "completed"
    CANCELLED = "cancelled"


@dataclass
class CostRecord:
    """Historical cost record."""
    project_id: str
    project_name: str
    work_item_code: str
    quantity: float
    estimated_cost: float
    actual_cost: float
    variance: float
    variance_percent: float
    completion_date: date
    notes: str = ""


@dataclass
class ProjectCostSummary:
    """Project cost summary."""
    project_id: str
    project_name: str
    project_type: str
    location: str
    status: ProjectStatus
    estimated_total: float
    actual_total: float
    variance: float
    variance_percent: float
    start_date: date
    completion_date: Optional[date]
    item_count: int


class CWICRHistoricalCost:
    """Track historical costs using CWICR data."""

    def __init__(self, cwicr_data: pd.DataFrame = None):
        self.cwicr = cwicr_data
        self._projects: Dict[str, ProjectCostSummary] = {}
        self._records: List[CostRecord] = []

        if cwicr_data is not None:
            self._index_cwicr()

    def _index_cwicr(self):
        """Index CWICR data."""
        if 'work_item_code' in self.cwicr.columns:
            self._cwicr_index = self.cwicr.set_index('work_item_code')
        else:
            self._cwicr_index = None

    def add_project(self,
                    project_id: str,
                    project_name: str,
                    project_type: str,
                    location: str,
                    estimated_total: float,
                    start_date: date) -> str:
        """Add new project to historical database."""

        summary = ProjectCostSummary(
            project_id=project_id,
            project_name=project_name,
            project_type=project_type,
            location=location,
            status=ProjectStatus.ESTIMATED,
            estimated_total=estimated_total,
            actual_total=0,
            variance=0,
            variance_percent=0,
            start_date=start_date,
            completion_date=None,
            item_count=0
        )

        self._projects[project_id] = summary
        return project_id

    def record_actual_cost(self,
                           project_id: str,
                           work_item_code: str,
                           quantity: float,
                           actual_cost: float,
                           completion_date: date = None,
                           notes: str = "") -> CostRecord:
        """Record actual cost for work item."""

        # Get estimated cost from CWICR
        estimated_unit_cost = 0
        if self._cwicr_index is not None and work_item_code in self._cwicr_index.index:
            item = self._cwicr_index.loc[work_item_code]
            labor = float(item.get('labor_cost', 0) or 0)
            material = float(item.get('material_cost', 0) or 0)
            equipment = float(item.get('equipment_cost', 0) or 0)
            estimated_unit_cost = labor + material + equipment

        estimated_cost = estimated_unit_cost * quantity
        variance = actual_cost - estimated_cost
        variance_pct = (variance / estimated_cost * 100) if estimated_cost > 0 else 0

        record = CostRecord(
            project_id=project_id,
            project_name=self._projects.get(project_id, {}).project_name if project_id in self._projects else "",
            work_item_code=work_item_code,
            quantity=quantity,
            estimated_cost=round(estimated_cost, 2),
            actual_cost=round(actual_cost, 2),
            variance=round(variance, 2),
            variance_percent=round(variance_pct, 1),
            completion_date=completion_date or date.today(),
            notes=notes
        )

        self._records.append(record)

        # Update project summary
        if project_id in self._projects:
            proj = self._projects[project_id]
            proj.actual_total += actual_cost
            proj.variance = proj.actual_total - proj.estimated_total
            proj.variance_percent = (proj.variance / proj.estimated_total * 100) if proj.estimated_total > 0 else 0
            proj.item_count += 1
            proj.status = ProjectStatus.IN_PROGRESS

        return record

    def complete_project(self, project_id: str, completion_date: date = None):
        """Mark project as completed."""
        if project_id in self._projects:
            self._projects[project_id].status = ProjectStatus.COMPLETED
            self._projects[project_id].completion_date = completion_date or date.today()

    def get_work_item_history(self, work_item_code: str) -> Dict[str, Any]:
        """Get historical data for specific work item."""

        records = [r for r in self._records if r.work_item_code == work_item_code]

        if not records:
            return {'work_item_code': work_item_code, 'records': 0}

        variances = [r.variance_percent for r in records]
        actual_costs = [r.actual_cost / r.quantity if r.quantity > 0 else 0 for r in records]

        return {
            'work_item_code': work_item_code,
            'records': len(records),
            'average_variance_pct': round(np.mean(variances), 1),
            'variance_std': round(np.std(variances), 1),
            'average_actual_unit_cost': round(np.mean(actual_costs), 2),
            'min_actual_unit_cost': round(min(actual_costs), 2),
            'max_actual_unit_cost': round(max(actual_costs), 2),
            'projects': list(set(r.project_id for r in records)),
            'trend': 'increasing' if len(records) > 2 and actual_costs[-1] > actual_costs[0] else 'stable'
        }

    def get_accuracy_metrics(self) -> Dict[str, Any]:
        """Calculate overall estimating accuracy metrics."""

        if not self._records:
            return {}

        variances = [r.variance_percent for r in self._records]

        # Accuracy by category
        by_category = {}
        for record in self._records:
            category = record.work_item_code.split('-')[0] if '-' in record.work_item_code else 'Other'
            if category not in by_category:
                by_category[category] = []
            by_category[category].append(record.variance_percent)

        category_accuracy = {
            cat: {
                'average_variance': round(np.mean(vals), 1),
                'count': len(vals)
            }
            for cat, vals in by_category.items()
        }

        return {
            'total_records': len(self._records),
            'average_variance_pct': round(np.mean(variances), 1),
            'variance_std': round(np.std(variances), 1),
            'within_5pct': sum(1 for v in variances if abs(v) <= 5) / len(variances) * 100,
            'within_10pct': sum(1 for v in variances if abs(v) <= 10) / len(variances) * 100,
            'overestimated_pct': sum(1 for v in variances if v < 0) / len(variances) * 100,
            'underestimated_pct': sum(1 for v in variances if v > 0) / len(variances) * 100,
            'by_category': category_accuracy
        }

    def suggest_adjustment_factors(self) -> Dict[str, float]:
        """Suggest adjustment