crushdataai 1.0.0

package/assets/antigravity/data-analyst.md

@@ -0,0 +1,95 @@
+---
+description: CrushData AI - Data Analyst workflow for structured analysis with validation
+---
+
+# CrushData AI - Data Analyst Workflow
+
+A data analyst intelligence workflow that guides you through structured, professional data analysis.
+
+## When to Use
+
+Use this workflow when user requests:
+- Data analysis, EDA, ad-hoc queries
+- Dashboard or report creation
+- Metrics calculation (MRR, churn, conversion, etc.)
+- Cohort, funnel, or A/B test analysis
+- Data cleaning or profiling
+
+---
+
+## Step 1: Discovery Protocol (MANDATORY)
+
+Before writing any code, ask the user:
+
+1. **Business Context**: What question should this analysis answer? Who is the audience?
+2. **Data Context**: Which tables contain the data? What time range?
+3. **Metric Definitions**: How does YOUR company define the key metrics? Any filters?
+
+---
+
+## Step 2: Search Knowledge Base
+
+// turbo
+```bash
+python3 .agent/workflows/../.shared/data-analyst/scripts/search.py "<query>" --domain <domain>
+```
+
+**Domains:**
+- `workflow` - Step-by-step analysis process
+- `metric` - Metric definitions and formulas
+- `chart` - Visualization recommendations
+- `sql` - SQL patterns (window functions, cohorts)
+- `python` - pandas/polars snippets
+- `validation` - Common mistakes to avoid
+
+**Industry search:**
+// turbo
+```bash
+python3 .shared/data-analyst/scripts/search.py "<query>" --industry saas|ecommerce|finance|marketing
+```
+
+---
+
+## Step 3: Data Profiling (MANDATORY)
+
+Before analysis, run profiling and report to user:
+
+```python
+print(f"Shape: {df.shape}")
+print(f"Date range: {df['date'].min()} to {df['date'].max()}")
+print(f"Missing values:\n{df.isnull().sum()}")
+```
+
+Ask: "I found X rows, Y users, dates from A to B. Does this match expectation?"
+
+---
+
+## Step 4: Execute with Validation
+
+**Before JOINs:** Run on 100 rows first. Ask if row count change is expected.
+
+**Before Aggregations:** Check for duplicates. Ask if totals seem reasonable.
+
+**Before Delivery:** Compare to benchmarks. Present for user validation.
+
+---
+
+## Search Examples
+
+| Analysis | Command |
+|----------|---------|
+| EDA workflow | `search.py "EDA" --domain workflow` |
+| Churn metrics | `search.py "churn" --domain metric` |
+| Cohort SQL | `search.py "cohort" --domain sql` |
+| SaaS metrics | `search.py "MRR" --industry saas` |
+| Chart selection | `search.py "time series" --domain chart` |
+
+---
+
+## Pre-Delivery Checklist
+
+- [ ] Business question answered
+- [ ] Data profiled and validated
+- [ ] Metric definitions confirmed with user
+- [ ] Sanity checks passed
+- [ ] Assumptions documented

package/assets/claude/SKILL.md

@@ -0,0 +1,145 @@
+# CrushData AI - Data Analyst Skill
+
+A data analyst intelligence skill that guides you through structured, professional data analysis workflows.
+
+---
+
+## How to Use This Skill
+
+When user requests data analysis work (analyze, query, dashboard, metrics, EDA, cohort, funnel, A/B test), follow this workflow:
+
+### Step 1: Discovery Protocol (MANDATORY)
+
+**Before writing any code, ask the user:**
+
+```
+## Discovery Questions
+
+1. **Business Context**
+   - What business question should this analysis answer?
+   - Who is the audience? (Executive, Analyst, Engineer)
+   - What action will this analysis inform?
+
+2. **Data Context**
+   - Which tables/databases contain the relevant data?
+   - What time range should I analyze?
+   - Any known data quality issues?
+
+3. **Metric Definitions**
+   - How does YOUR company define the key metrics?
+   - Any filters to apply? (exclude test users, internal accounts?)
+   - What timezone should I use for dates?
+```
+
+### Step 2: Search Relevant Domains
+
+Use `search.py` to gather comprehensive information:
+
+```bash
+python3 .claude/skills/data-analyst/scripts/search.py "<query>" --domain <domain> [-n 3]
+```
+
+**Available domains:**
+| Domain | Use Case |
+|--------|----------|
+| `workflow` | Step-by-step analysis process |
+| `metric` | Metric definitions and formulas |
+| `chart` | Visualization recommendations |
+| `cleaning` | Data quality patterns |
+| `sql` | SQL patterns (window functions, cohorts) |
+| `python` | pandas/polars code snippets |
+| `database` | PostgreSQL, BigQuery, Snowflake tips |
+| `report` | Dashboard UX guidelines |
+| `validation` | Common mistakes to avoid |
+
+**Industry-specific search:**
+```bash
+python3 .claude/skills/data-analyst/scripts/search.py "<query>" --industry saas|ecommerce|finance|marketing
+```
+
+**Recommended search order:**
+1. `workflow` - Get the step-by-step process for this analysis type
+2. `metric` or `--industry` - Get relevant metric definitions
+3. `sql` or `python` - Get code patterns for implementation
+4. `chart` - Get visualization recommendations
+5. `validation` - Check for common mistakes to avoid
+
+### Step 3: Data Profiling (MANDATORY Before Analysis)
+
+Before any analysis, run profiling:
+
+**Python:**
+```python
+print(f"Shape: {df.shape}")
+print(f"Date range: {df['date'].min()} to {df['date'].max()}")
+print(f"Missing values:\n{df.isnull().sum()}")
+print(f"Sample:\n{df.sample(5)}")
+```
+
+**SQL:**
+```sql
+SELECT 
+    COUNT(*) as total_rows,
+    COUNT(DISTINCT user_id) as unique_users,
+    MIN(date) as min_date,
+    MAX(date) as max_date
+FROM table;
+```
+
+**Report findings to user before proceeding:**
+> "I found X rows, Y unique users, date range from A to B. Does this match your expectation?"
+
+### Step 4: Execute Analysis with Validation
+
+**Before JOINs:**
+- Run on 100 rows first
+- Check: Did row count change unexpectedly?
+- Ask: "The join produced X rows from Y. Expected?"
+
+**Before Aggregations:**
+- Check for duplicates that could inflate sums
+- Verify granularity: "Is this one row per user per day?"
+- Ask: "Total = $X. Does this seem reasonable?"
+
+**Before Delivery:**
+- Sanity check order of magnitude
+- Compare to benchmark or prior period
+- Present for user validation before finalizing
+
+---
+
+## Workflow Reference
+
+| Analysis Type | Search Command |
+|--------------|----------------|
+| EDA | `--domain workflow` query "exploratory data analysis" |
+| Dashboard | `--domain workflow` query "dashboard creation" |
+| A/B Test | `--domain workflow` query "ab test" |
+| Cohort | `--domain workflow` query "cohort analysis" |
+| Funnel | `--domain workflow` query "funnel analysis" |
+| Time Series | `--domain workflow` query "time series" |
+| Segmentation | `--domain workflow` query "customer segmentation" |
+| Data Cleaning | `--domain workflow` query "data cleaning" |
+
+---
+
+## Common Rules
+
+1. **Always ask before assuming** - Metric definitions vary by company
+2. **Profile data first** - Never aggregate without understanding the data
+3. **Validate results** - Check totals, compare to benchmarks
+4. **Document assumptions** - State what filters and definitions you used
+5. **Show your work** - Explain the logic behind complex queries
+
+---
+
+## Pre-Delivery Checklist
+
+Before presenting final results:
+
+- [ ] Confirmed business question is answered
+- [ ] Data was profiled and validated
+- [ ] Metric definitions match user's expectations
+- [ ] Sanity checks pass (order of magnitude, trends, etc.)
+- [ ] Visualizations follow best practices (search `--domain chart`)
+- [ ] Assumptions and filters are documented

package/assets/copilot/data-analyst.prompt.md

@@ -0,0 +1,40 @@
+---
+mode: agent
+description: CrushData AI - Data analyst intelligence for structured analysis workflows
+tools: ['codebase', 'terminal', 'file-operations']
+---
+
+# CrushData AI - Data Analyst
+
+Guide structured, professional data analysis with validation.
+
+## When to Activate
+User requests: data analysis, dashboards, metrics, EDA, cohort, funnel, A/B tests
+
+## Workflow
+
+### 1. Discovery (MANDATORY)
+Before coding, ask:
+- What business question should this answer?
+- Which tables contain the data?
+- How does YOUR company define key metrics?
+
+### 2. Search Knowledge Base
+```bash
+python3 .github/prompts/../.shared/data-analyst/scripts/search.py "<query>" --domain <domain>
+```
+
+Domains: `workflow`, `metric`, `chart`, `sql`, `python`, `validation`
+Industry: `--industry saas|ecommerce|finance|marketing`
+
+### 3. Profile Data
+```python
+print(f"Shape: {df.shape}, Dates: {df['date'].min()} to {df['date'].max()}")
+```
+Report and confirm before proceeding.
+
+### 4. Validate
+- Verify JOINs
+- Check totals
+- Compare benchmarks
+- User validation

package/assets/cursor/data-analyst.md

@@ -0,0 +1,50 @@
+# CrushData AI - Data Analyst Command
+
+A data analyst intelligence command that guides you through structured, professional data analysis.
+
+## When to Use
+
+Activate this command when user requests data analysis, dashboards, metrics, EDA, cohort/funnel analysis, or A/B testing.
+
+---
+
+## Workflow
+
+### 1. Discovery (MANDATORY)
+Before coding, ask:
+- What business question should this answer?
+- Which tables contain the data?
+- How does YOUR company define the key metrics?
+
+### 2. Search Knowledge Base
+```bash
+python3 .cursor/commands/../.shared/data-analyst/scripts/search.py "<query>" --domain <domain>
+```
+
+Domains: `workflow`, `metric`, `chart`, `sql`, `python`, `validation`
+
+Industry: `--industry saas|ecommerce|finance|marketing`
+
+### 3. Data Profiling (MANDATORY)
+```python
+print(f"Shape: {df.shape}, Date range: {df['date'].min()} to {df['date'].max()}")
+```
+Report findings and ask user for confirmation.
+
+### 4. Validate Before Delivery
+- Check JOINs don't multiply rows unexpectedly
+- Verify totals seem reasonable
+- Compare to benchmarks
+- Present for user validation
+
+---
+
+## Quick Reference
+
+| Analysis | Search Command |
+|----------|---------------|
+| EDA | `--domain workflow` "EDA" |
+| Metrics | `--domain metric` "churn" |
+| SQL patterns | `--domain sql` "cohort" |
+| Charts | `--domain chart` "time series" |
+| Mistakes | `--domain validation` "duplicate" |

package/assets/kiro/data-analyst.md

@@ -0,0 +1,37 @@
+# CrushData AI - Data Analyst Steering
+
+## Purpose
+Guide structured, professional data analysis with validation protocols.
+
+## When to Activate
+User requests: data analysis, dashboards, metrics, EDA, cohort, funnel, A/B tests
+
+## Required Behaviors
+
+### 1. Always Ask First
+Before writing code, gather:
+- Business question to answer
+- Data tables/sources
+- Company-specific metric definitions
+- Time range and filters
+
+### 2. Search Before Implementing
+```bash
+python3 .kiro/steering/../.shared/data-analyst/scripts/search.py "<query>" --domain <domain>
+```
+
+Available domains: workflow, metric, chart, sql, python, database, validation
+
+### 3. Profile Data Before Analysis
+Run and report:
+- Row counts and date ranges
+- Missing values
+- Sample data
+
+Ask: "Does this match your expectation?"
+
+### 4. Validate Before Delivery
+- Sanity check totals
+- Compare to benchmarks
+- Document assumptions
+- Present for user confirmation

package/assets/shared/charts.csv

@@ -0,0 +1,31 @@
+Chart Type,Best For,Data Type,Comparison Type,Python Code,Color Guidance,Accessibility,Dashboard Tip
+Line Chart,Trends over time and continuous data,Time-series,Trend,"plt.plot(df['date'], df['value']); plt.xlabel('Date'); plt.ylabel('Value')","Sequential blue/green for single metric; categorical for multiple series","Add markers for key data points; use sufficient line thickness","Place in middle section for trend visibility"
+Bar Chart,Comparing categories or rankings,Categorical,Ranking Comparison,"plt.bar(df['category'], df['value']); plt.xticks(rotation=45)","Single color for one series; categorical for grouped","Ensure sufficient contrast between bars; label values directly","Use horizontal layout if labels are long"
+Horizontal Bar Chart,Ranking with long labels,Categorical,Ranking,"plt.barh(df['category'], df['value'])","Single sequential color from light to dark by value","Order bars by value for easy scanning","Great for top-N lists and leaderboards"
+Stacked Bar Chart,Part-to-whole over categories,Categorical,Composition,"df.plot(kind='bar', stacked=True)","Use distinct colors for each segment; max 5-6 segments","Include legend; consider labels on large segments","Good for showing composition across time periods"
+Grouped Bar Chart,Comparing multiple series by category,Categorical,Comparison,"df.plot(kind='bar', position='dodge')","Categorical palette with clear distinction between groups","Limit to 3-4 groups per category; use legend","Best for A/B comparisons or time period comparisons"
+Pie Chart,Simple part-to-whole (max 5 segments),Categorical,Composition,"plt.pie(df['value'], labels=df['category'], autopct='%1.1f%%')","High contrast between adjacent segments","Limit to 5 segments; order by size; include percentages","Avoid in dashboards - use donut or bar instead"
+Donut Chart,Part-to-whole with center metric,Categorical,Composition,"plt.pie(df['value'], wedgeprops={'width': 0.4})","High contrast colors; use brand colors if relevant","Include total or key metric in center","Good for single KPI with breakdown"
+Area Chart,Trends with volume emphasis,Time-series,Trend Volume,"plt.fill_between(df['date'], df['value'], alpha=0.3)","Light fill with darker line; sequential colors","Ensure baseline is visible; use transparency","Shows magnitude over time better than line"
+Stacked Area Chart,Composition over time,Time-series,Composition Trend,"df.plot(kind='area', stacked=True)","Distinct colors for each layer; limit layers to 5","Consider 100% stacked for proportion focus","Good for market share or category mix over time"
+Scatter Plot,Relationship between two variables,Numerical,Correlation,"plt.scatter(df['x'], df['y'])","Single color with alpha for density; color by category if needed","Add trendline for correlation; include R-squared","Use for identifying outliers and patterns"
+Bubble Chart,Three-variable relationships,Numerical,Correlation Size,"plt.scatter(df['x'], df['y'], s=df['size']*100)","Color by category if applicable; size legend essential","Ensure bubbles don't overlap too much","Include size legend; limit to important points"
+Heatmap,Correlations or matrix data,Numerical Matrix,Distribution,"sns.heatmap(df.corr(), annot=True, cmap='RdBu_r')","Diverging palette for correlation (-1 to 1); sequential for counts","Include value annotations; use colorblind-safe palette","Perfect for cohort retention tables"
+Histogram,Distribution of single variable,Numerical,Distribution,"plt.hist(df['value'], bins=30)","Single color; consider outlier highlighting","Include mean/median line; label bin count","Use to understand data distribution before analysis"
+Box Plot,Distribution comparison,Numerical,Distribution Comparison,"sns.boxplot(data=df, x='category', y='value')","One color per category; highlight outliers","Explain quartile meanings; include n count","Great for comparing distributions across groups"
+Violin Plot,Distribution with density,Numerical,Distribution,"sns.violinplot(data=df, x='category', y='value')","Paired colors for split violins; sequential otherwise","More intuitive than box plots for some users","Good for showing bimodal distributions"
+Funnel Chart,Sequential step conversion,Categorical,Drop-off,"import plotly.express as px; px.funnel(df, x='count', y='stage')","Blues from dark to light (top to bottom); or brand colors","Label conversion percentages between stages","Essential for showing conversion drop-off"
+Waterfall Chart,Cumulative effect of values,Categorical,Contribution,"Use plotly or custom matplotlib with positive/negative coloring","Green for positive; red for negative; gray for subtotals","Start with total; show increases and decreases clearly","Great for bridge charts (start to end explanation)"
+Gauge Chart,Single KPI with target,Single Value,Target,"Use plotly Indicator or custom graphic","Green/yellow/red zones based on targets","Include actual value and target","Use sparingly - one per major KPI"
+KPI Card,Single important metric,Single Value,Status,"Text display with conditional formatting","Color based on performance (green/amber/red)","Large font; include trend arrow and context","Top of dashboard for most important metrics"
+Sparkline,Compact trend indicator,Time-series,Trend,"Line chart rendered small without axes","Single color; consistent across dashboard","May be too small for some users","Great alongside KPI cards to show trend"
+Table,Detailed data viewing,Multi-dimensional,Detail,"df.style.format() with conditional formatting","Alternate row colors; highlight important values","Ensure sufficient contrast; limit columns","Place at bottom of dashboard for drill-down"
+Pivot Table,Cross-tabulation analysis,Multi-dimensional,Comparison,"pd.pivot_table(df, values='metric', index='row', columns='col')","Heatmap coloring on values if applicable","Include row/column totals","Good for exploration; use charts for communication"
+Treemap,Hierarchical part-to-whole,Hierarchical,Composition,"import plotly.express as px; px.treemap(df, path=['parent', 'child'], values='value')","Distinct colors per category; size shows proportion","Include value labels on large segments","Good for budget/allocation visualization"
+Sankey Diagram,Flow between categories,Flow,Flow,"Use plotly Sankey for flow visualization","Distinct colors per source node","Limit to 5-10 nodes for readability","Perfect for showing customer journeys"
+Radar/Spider Chart,Multi-variable comparison,Multi-dimensional,Profile,"Use matplotlib radar chart or plotly","One color per entity being compared","Include reference lines; limit to 5-8 axes","Good for segment profiles or competitive analysis"
+Geographic Map,Location-based data,Geographic,Distribution,"Use folium or plotly for choropleth maps","Sequential color scale for values","Ensure color scale is clear; include legend","Use for regional performance comparisons"
+Calendar Heatmap,Activity over time,Time-series,Pattern,"Use calplot or custom heatmap by day","Sequential palette; highlight weekends differently","Label axes clearly; include color legend","Good for showing seasonal patterns"
+Combination Chart,Mixed data types,Mixed,Correlation Trend,"Use secondary y-axis: ax2 = ax1.twinx()","Distinct colors for each series; clear legend","Ensure both scales are readable","Use when showing related but different units"
+Small Multiples,Comparison across categories,Multi-dimensional,Comparison,"Use facet plot: sns.FacetGrid(df, col='category')","Consistent scale and colors across all charts","Keep individual charts simple","Great for comparing patterns across segments"
+Bullet Chart,KPI vs target and comparison,Single Value,Target Comparison,"Use plotly for bullet chart implementation","Gray for comparison; bars for actual/target","Include labels for all elements","Compact alternative to gauge for multiple KPIs"

package/assets/shared/cleaning.csv

@@ -0,0 +1,21 @@
+Issue Type,Detection Method,Solution,Python Code,SQL Code,Impact
+Missing Values,df.isnull().sum() and df.isnull().mean(),"Drop rows, impute with mean/median/mode, forward fill, or flag with indicator","df['col'].fillna(df['col'].median(), inplace=True) or df.dropna(subset=['required_col'])","COALESCE(column, 0) or WHERE column IS NOT NULL","Missing data can skew aggregations, break joins, and cause errors"
+Duplicate Rows,df.duplicated().sum() and df[df.duplicated()],"Remove exact duplicates or dedupe by key keeping latest","df.drop_duplicates(subset=['id'], keep='last', inplace=True)","WITH ranked AS (SELECT *, ROW_NUMBER() OVER (PARTITION BY id ORDER BY updated_at DESC) as rn FROM table) SELECT * FROM ranked WHERE rn = 1","Duplicates inflate counts, sums, and cause join multiplication"
+Outliers,df.describe() and boxplot IQR method,"Remove, cap/Winsorize, or investigate - depends on domain","Q1, Q3 = df['col'].quantile([0.25, 0.75]); IQR = Q3-Q1; df = df[(df['col'] >= Q1-1.5*IQR) & (df['col'] <= Q3+1.5*IQR)]","WHERE value BETWEEN (SELECT PERCENTILE_CONT(0.25) - 1.5*IQR) AND (SELECT PERCENTILE_CONT(0.75) + 1.5*IQR)","Outliers can dominate averages and distort visualizations"
+Data Type Mismatch,df.dtypes and df['col'].apply(type).value_counts(),"Convert to correct type with error handling","df['date'] = pd.to_datetime(df['date'], errors='coerce'); df['amount'] = pd.to_numeric(df['amount'], errors='coerce')","CAST(column AS DATE) or TRY_CAST for safe conversion","Wrong types cause sorting, filtering, and aggregation errors"
+Inconsistent Date Formats,df['date'].str.contains(pattern).value_counts(),"Standardize to ISO format YYYY-MM-DD","df['date'] = pd.to_datetime(df['date'], format='mixed').dt.strftime('%Y-%m-%d')","TO_DATE(date_string, 'format pattern')","Inconsistent dates cause parsing errors and incorrect sorting"
+Leading/Trailing Whitespace,df['col'].str.len() vs df['col'].str.strip().str.len(),"Strip whitespace from string columns","df['col'] = df['col'].str.strip()","TRIM(column)","Whitespace causes join failures and lookup misses"
+Case Inconsistency,df['col'].str.lower().nunique() vs df['col'].nunique(),"Standardize to lowercase or title case","df['col'] = df['col'].str.lower() or .str.title()","LOWER(column) or UPPER(column)","Case differences cause grouping errors and aggregation issues"
+Invalid Categories,df['category'].isin(valid_list).value_counts(),"Map invalid values or flag/remove","df['category'] = df['category'].replace({'invalid': 'Unknown'}); df = df[df['category'].isin(valid_list)]","CASE WHEN category IN ('valid1', 'valid2') THEN category ELSE 'Other' END","Invalid categories skew analysis and break filters"
+Negative Values Where Impossible,df[df['col'] < 0].count(),"Flag, remove, or convert to absolute value","df = df[df['quantity'] >= 0] or df['quantity'] = df['quantity'].abs()","WHERE quantity >= 0 or ABS(quantity)","Negative quantities/prices indicate data entry errors"
+Future Dates in Historical Data,df[df['date'] > pd.Timestamp.today()],"Remove or flag future-dated records","df = df[df['date'] <= pd.Timestamp.today()]","WHERE date <= CURRENT_DATE","Future dates indicate data pipeline or entry errors"
+Zero Division Risk,df[df['denominator'] == 0].count(),"Handle zeros before division with NULLIF or fillna","df['ratio'] = df['numerator'] / df['denominator'].replace(0, np.nan)","numerator / NULLIF(denominator, 0)","Division by zero causes errors or inf values"
+Encoding Issues,df['col'].str.contains('[^\x00-\x7F]'),"Fix encoding or remove special characters","df['col'] = df['col'].str.encode('ascii', 'ignore').str.decode('ascii')","REGEXP_REPLACE(col, '[^[:ascii:]]', '')","Encoding issues cause display and processing errors"
+Null vs Zero Ambiguity,df['col'].isin([0, None]).value_counts(),"Decide semantic meaning: is 0 different from null?","Document decision: df['col'] = df['col'].fillna(0) # if semantically equivalent","Add explicit flag: CASE WHEN col IS NULL THEN 'Unknown' ELSE 'Known' END","Confusing null with zero leads to calculation errors"
+Data Entry Typos,df['name'].str.lower().value_counts() looking for similar values,"Use fuzzy matching to identify and merge typos","from fuzzywuzzy import fuzz; identify similar strings","Use pg_trgm or Levenshtein distance functions","Typos split metrics that should be grouped together"
+Orphan Records,df.merge(reference_df, how='left', indicator=True).query('_merge == \"left_only\"'),"Remove orphans or add to reference table","df = df[df['foreign_key'].isin(reference_df['id'])]","WHERE foreign_key IN (SELECT id FROM reference_table)","Orphan records indicate referential integrity issues"
+Mixed Numeric Formats,df['col'].str.contains(r'[\$,€%]'),"Extract numeric values removing currency/percent symbols","df['amount'] = df['amount'].str.replace('[$,]', '', regex=True).astype(float)","CAST(REPLACE(REPLACE(col, '$', ''), ',', '') AS DECIMAL)","Mixed formats prevent numeric operations"
+Boolean Inconsistency,df['flag'].unique() showing mixed True/False/1/0/Yes/No,"Standardize to consistent boolean representation","df['flag'] = df['flag'].map({'Yes': True, 'No': False, 1: True, 0: False})","CASE WHEN flag IN ('Yes', 'Y', '1', 'true') THEN TRUE ELSE FALSE END","Inconsistent booleans cause filtering errors"
+Data Freshness,df['updated_at'].max() vs expected freshness,"Alert if data is stale beyond threshold","assert (pd.Timestamp.today() - df['updated_at'].max()).days < 1, 'Data is stale'","WHERE updated_at >= CURRENT_DATE - INTERVAL '1 day'","Stale data leads to incorrect analysis and decisions"
+Cardinality Changes,Compare df['col'].nunique() to historical baseline,"Alert if cardinality changes unexpectedly","assert df['category'].nunique() == expected_count, f'Expected {expected_count} categories'","SELECT COUNT(DISTINCT col) and compare to metadata","New or missing categories indicate upstream issues"
+Range Violations,df[~df['col'].between(min_val, max_val)],"Flag or remove out-of-range values","df = df[df['age'].between(0, 120)]","WHERE age BETWEEN 0 AND 120","Out-of-range values indicate data quality issues"

package/assets/shared/data/charts.csv

@@ -0,0 +1,31 @@
+Chart Type,Best For,Data Type,Comparison Type,Python Code,Color Guidance,Accessibility,Dashboard Tip
+Line Chart,Trends over time and continuous data,Time-series,Trend,"plt.plot(df['date'], df['value']); plt.xlabel('Date'); plt.ylabel('Value')","Sequential blue/green for single metric; categorical for multiple series","Add markers for key data points; use sufficient line thickness","Place in middle section for trend visibility"
+Bar Chart,Comparing categories or rankings,Categorical,Ranking Comparison,"plt.bar(df['category'], df['value']); plt.xticks(rotation=45)","Single color for one series; categorical for grouped","Ensure sufficient contrast between bars; label values directly","Use horizontal layout if labels are long"
+Horizontal Bar Chart,Ranking with long labels,Categorical,Ranking,"plt.barh(df['category'], df['value'])","Single sequential color from light to dark by value","Order bars by value for easy scanning","Great for top-N lists and leaderboards"
+Stacked Bar Chart,Part-to-whole over categories,Categorical,Composition,"df.plot(kind='bar', stacked=True)","Use distinct colors for each segment; max 5-6 segments","Include legend; consider labels on large segments","Good for showing composition across time periods"
+Grouped Bar Chart,Comparing multiple series by category,Categorical,Comparison,"df.plot(kind='bar', position='dodge')","Categorical palette with clear distinction between groups","Limit to 3-4 groups per category; use legend","Best for A/B comparisons or time period comparisons"
+Pie Chart,Simple part-to-whole (max 5 segments),Categorical,Composition,"plt.pie(df['value'], labels=df['category'], autopct='%1.1f%%')","High contrast between adjacent segments","Limit to 5 segments; order by size; include percentages","Avoid in dashboards - use donut or bar instead"
+Donut Chart,Part-to-whole with center metric,Categorical,Composition,"plt.pie(df['value'], wedgeprops={'width': 0.4})","High contrast colors; use brand colors if relevant","Include total or key metric in center","Good for single KPI with breakdown"
+Area Chart,Trends with volume emphasis,Time-series,Trend Volume,"plt.fill_between(df['date'], df['value'], alpha=0.3)","Light fill with darker line; sequential colors","Ensure baseline is visible; use transparency","Shows magnitude over time better than line"
+Stacked Area Chart,Composition over time,Time-series,Composition Trend,"df.plot(kind='area', stacked=True)","Distinct colors for each layer; limit layers to 5","Consider 100% stacked for proportion focus","Good for market share or category mix over time"
+Scatter Plot,Relationship between two variables,Numerical,Correlation,"plt.scatter(df['x'], df['y'])","Single color with alpha for density; color by category if needed","Add trendline for correlation; include R-squared","Use for identifying outliers and patterns"
+Bubble Chart,Three-variable relationships,Numerical,Correlation Size,"plt.scatter(df['x'], df['y'], s=df['size']*100)","Color by category if applicable; size legend essential","Ensure bubbles don't overlap too much","Include size legend; limit to important points"
+Heatmap,Correlations or matrix data,Numerical Matrix,Distribution,"sns.heatmap(df.corr(), annot=True, cmap='RdBu_r')","Diverging palette for correlation (-1 to 1); sequential for counts","Include value annotations; use colorblind-safe palette","Perfect for cohort retention tables"
+Histogram,Distribution of single variable,Numerical,Distribution,"plt.hist(df['value'], bins=30)","Single color; consider outlier highlighting","Include mean/median line; label bin count","Use to understand data distribution before analysis"
+Box Plot,Distribution comparison,Numerical,Distribution Comparison,"sns.boxplot(data=df, x='category', y='value')","One color per category; highlight outliers","Explain quartile meanings; include n count","Great for comparing distributions across groups"
+Violin Plot,Distribution with density,Numerical,Distribution,"sns.violinplot(data=df, x='category', y='value')","Paired colors for split violins; sequential otherwise","More intuitive than box plots for some users","Good for showing bimodal distributions"
+Funnel Chart,Sequential step conversion,Categorical,Drop-off,"import plotly.express as px; px.funnel(df, x='count', y='stage')","Blues from dark to light (top to bottom); or brand colors","Label conversion percentages between stages","Essential for showing conversion drop-off"
+Waterfall Chart,Cumulative effect of values,Categorical,Contribution,"Use plotly or custom matplotlib with positive/negative coloring","Green for positive; red for negative; gray for subtotals","Start with total; show increases and decreases clearly","Great for bridge charts (start to end explanation)"
+Gauge Chart,Single KPI with target,Single Value,Target,"Use plotly Indicator or custom graphic","Green/yellow/red zones based on targets","Include actual value and target","Use sparingly - one per major KPI"
+KPI Card,Single important metric,Single Value,Status,"Text display with conditional formatting","Color based on performance (green/amber/red)","Large font; include trend arrow and context","Top of dashboard for most important metrics"
+Sparkline,Compact trend indicator,Time-series,Trend,"Line chart rendered small without axes","Single color; consistent across dashboard","May be too small for some users","Great alongside KPI cards to show trend"
+Table,Detailed data viewing,Multi-dimensional,Detail,"df.style.format() with conditional formatting","Alternate row colors; highlight important values","Ensure sufficient contrast; limit columns","Place at bottom of dashboard for drill-down"
+Pivot Table,Cross-tabulation analysis,Multi-dimensional,Comparison,"pd.pivot_table(df, values='metric', index='row', columns='col')","Heatmap coloring on values if applicable","Include row/column totals","Good for exploration; use charts for communication"
+Treemap,Hierarchical part-to-whole,Hierarchical,Composition,"import plotly.express as px; px.treemap(df, path=['parent', 'child'], values='value')","Distinct colors per category; size shows proportion","Include value labels on large segments","Good for budget/allocation visualization"
+Sankey Diagram,Flow between categories,Flow,Flow,"Use plotly Sankey for flow visualization","Distinct colors per source node","Limit to 5-10 nodes for readability","Perfect for showing customer journeys"
+Radar/Spider Chart,Multi-variable comparison,Multi-dimensional,Profile,"Use matplotlib radar chart or plotly","One color per entity being compared","Include reference lines; limit to 5-8 axes","Good for segment profiles or competitive analysis"
+Geographic Map,Location-based data,Geographic,Distribution,"Use folium or plotly for choropleth maps","Sequential color scale for values","Ensure color scale is clear; include legend","Use for regional performance comparisons"
+Calendar Heatmap,Activity over time,Time-series,Pattern,"Use calplot or custom heatmap by day","Sequential palette; highlight weekends differently","Label axes clearly; include color legend","Good for showing seasonal patterns"
+Combination Chart,Mixed data types,Mixed,Correlation Trend,"Use secondary y-axis: ax2 = ax1.twinx()","Distinct colors for each series; clear legend","Ensure both scales are readable","Use when showing related but different units"
+Small Multiples,Comparison across categories,Multi-dimensional,Comparison,"Use facet plot: sns.FacetGrid(df, col='category')","Consistent scale and colors across all charts","Keep individual charts simple","Great for comparing patterns across segments"
+Bullet Chart,KPI vs target and comparison,Single Value,Target Comparison,"Use plotly for bullet chart implementation","Gray for comparison; bars for actual/target","Include labels for all elements","Compact alternative to gauge for multiple KPIs"

package/assets/shared/data/cleaning.csv

@@ -0,0 +1,21 @@
+Issue Type,Detection Method,Solution,Python Code,SQL Code,Impact
+Missing Values,df.isnull().sum() and df.isnull().mean(),"Drop rows, impute with mean/median/mode, forward fill, or flag with indicator","df['col'].fillna(df['col'].median(), inplace=True) or df.dropna(subset=['required_col'])","COALESCE(column, 0) or WHERE column IS NOT NULL","Missing data can skew aggregations, break joins, and cause errors"
+Duplicate Rows,df.duplicated().sum() and df[df.duplicated()],"Remove exact duplicates or dedupe by key keeping latest","df.drop_duplicates(subset=['id'], keep='last', inplace=True)","WITH ranked AS (SELECT *, ROW_NUMBER() OVER (PARTITION BY id ORDER BY updated_at DESC) as rn FROM table) SELECT * FROM ranked WHERE rn = 1","Duplicates inflate counts, sums, and cause join multiplication"
+Outliers,df.describe() and boxplot IQR method,"Remove, cap/Winsorize, or investigate - depends on domain","Q1, Q3 = df['col'].quantile([0.25, 0.75]); IQR = Q3-Q1; df = df[(df['col'] >= Q1-1.5*IQR) & (df['col'] <= Q3+1.5*IQR)]","WHERE value BETWEEN (SELECT PERCENTILE_CONT(0.25) - 1.5*IQR) AND (SELECT PERCENTILE_CONT(0.75) + 1.5*IQR)","Outliers can dominate averages and distort visualizations"
+Data Type Mismatch,df.dtypes and df['col'].apply(type).value_counts(),"Convert to correct type with error handling","df['date'] = pd.to_datetime(df['date'], errors='coerce'); df['amount'] = pd.to_numeric(df['amount'], errors='coerce')","CAST(column AS DATE) or TRY_CAST for safe conversion","Wrong types cause sorting, filtering, and aggregation errors"
+Inconsistent Date Formats,df['date'].str.contains(pattern).value_counts(),"Standardize to ISO format YYYY-MM-DD","df['date'] = pd.to_datetime(df['date'], format='mixed').dt.strftime('%Y-%m-%d')","TO_DATE(date_string, 'format pattern')","Inconsistent dates cause parsing errors and incorrect sorting"
+Leading/Trailing Whitespace,df['col'].str.len() vs df['col'].str.strip().str.len(),"Strip whitespace from string columns","df['col'] = df['col'].str.strip()","TRIM(column)","Whitespace causes join failures and lookup misses"
+Case Inconsistency,df['col'].str.lower().nunique() vs df['col'].nunique(),"Standardize to lowercase or title case","df['col'] = df['col'].str.lower() or .str.title()","LOWER(column) or UPPER(column)","Case differences cause grouping errors and aggregation issues"
+Invalid Categories,df['category'].isin(valid_list).value_counts(),"Map invalid values or flag/remove","df['category'] = df['category'].replace({'invalid': 'Unknown'}); df = df[df['category'].isin(valid_list)]","CASE WHEN category IN ('valid1', 'valid2') THEN category ELSE 'Other' END","Invalid categories skew analysis and break filters"
+Negative Values Where Impossible,df[df['col'] < 0].count(),"Flag, remove, or convert to absolute value","df = df[df['quantity'] >= 0] or df['quantity'] = df['quantity'].abs()","WHERE quantity >= 0 or ABS(quantity)","Negative quantities/prices indicate data entry errors"
+Future Dates in Historical Data,df[df['date'] > pd.Timestamp.today()],"Remove or flag future-dated records","df = df[df['date'] <= pd.Timestamp.today()]","WHERE date <= CURRENT_DATE","Future dates indicate data pipeline or entry errors"
+Zero Division Risk,df[df['denominator'] == 0].count(),"Handle zeros before division with NULLIF or fillna","df['ratio'] = df['numerator'] / df['denominator'].replace(0, np.nan)","numerator / NULLIF(denominator, 0)","Division by zero causes errors or inf values"
+Encoding Issues,df['col'].str.contains('[^\x00-\x7F]'),"Fix encoding or remove special characters","df['col'] = df['col'].str.encode('ascii', 'ignore').str.decode('ascii')","REGEXP_REPLACE(col, '[^[:ascii:]]', '')","Encoding issues cause display and processing errors"
+Null vs Zero Ambiguity,df['col'].isin([0, None]).value_counts(),"Decide semantic meaning: is 0 different from null?","Document decision: df['col'] = df['col'].fillna(0) # if semantically equivalent","Add explicit flag: CASE WHEN col IS NULL THEN 'Unknown' ELSE 'Known' END","Confusing null with zero leads to calculation errors"
+Data Entry Typos,df['name'].str.lower().value_counts() looking for similar values,"Use fuzzy matching to identify and merge typos","from fuzzywuzzy import fuzz; identify similar strings","Use pg_trgm or Levenshtein distance functions","Typos split metrics that should be grouped together"
+Orphan Records,df.merge(reference_df, how='left', indicator=True).query('_merge == \"left_only\"'),"Remove orphans or add to reference table","df = df[df['foreign_key'].isin(reference_df['id'])]","WHERE foreign_key IN (SELECT id FROM reference_table)","Orphan records indicate referential integrity issues"
+Mixed Numeric Formats,df['col'].str.contains(r'[\$,€%]'),"Extract numeric values removing currency/percent symbols","df['amount'] = df['amount'].str.replace('[$,]', '', regex=True).astype(float)","CAST(REPLACE(REPLACE(col, '$', ''), ',', '') AS DECIMAL)","Mixed formats prevent numeric operations"
+Boolean Inconsistency,df['flag'].unique() showing mixed True/False/1/0/Yes/No,"Standardize to consistent boolean representation","df['flag'] = df['flag'].map({'Yes': True, 'No': False, 1: True, 0: False})","CASE WHEN flag IN ('Yes', 'Y', '1', 'true') THEN TRUE ELSE FALSE END","Inconsistent booleans cause filtering errors"
+Data Freshness,df['updated_at'].max() vs expected freshness,"Alert if data is stale beyond threshold","assert (pd.Timestamp.today() - df['updated_at'].max()).days < 1, 'Data is stale'","WHERE updated_at >= CURRENT_DATE - INTERVAL '1 day'","Stale data leads to incorrect analysis and decisions"
+Cardinality Changes,Compare df['col'].nunique() to historical baseline,"Alert if cardinality changes unexpectedly","assert df['category'].nunique() == expected_count, f'Expected {expected_count} categories'","SELECT COUNT(DISTINCT col) and compare to metadata","New or missing categories indicate upstream issues"
+Range Violations,df[~df['col'].between(min_val, max_val)],"Flag or remove out-of-range values","df = df[df['age'].between(0, 120)]","WHERE age BETWEEN 0 AND 120","Out-of-range values indicate data quality issues"

package/assets/shared/data/databases.csv

@@ -0,0 +1,35 @@
+Database,Category,Guideline,Do,Don't,Code Example
+PostgreSQL,Connection,Use connection pooling for efficiency,"Use psycopg2 pool or SQLAlchemy with pool_size","Create new connection for each query","from sqlalchemy import create_engine; engine = create_engine('postgresql://...', pool_size=5)"
+PostgreSQL,Query,Use EXPLAIN ANALYZE for query tuning,"EXPLAIN ANALYZE your slow queries; check for seq scans","Optimize blindly without understanding execution plan","EXPLAIN ANALYZE SELECT * FROM orders WHERE date > '2024-01-01'"
+PostgreSQL,Indexing,Create indexes on filtered and joined columns,"CREATE INDEX idx_date ON orders(order_date)","Index every column; forget to ANALYZE after","CREATE INDEX CONCURRENTLY to avoid locking"
+PostgreSQL,Dates,Use date_trunc for time grouping,"date_trunc('month', order_date)","String manipulation on dates","SELECT date_trunc('day', ts) as day, COUNT(*) FROM events GROUP BY 1"
+PostgreSQL,Window,Use window functions for analytics,"OVER (PARTITION BY ... ORDER BY ...)","Self-joins for running totals","SUM(amount) OVER (ORDER BY date ROWS UNBOUNDED PRECEDING)"
+PostgreSQL,CTEs,Use CTEs for readable complex queries,"WITH step1 AS (...), step2 AS (...)","Deeply nested subqueries","WITH monthly AS (SELECT date_trunc('month', date) ...) SELECT * FROM monthly"
+BigQuery,Cost,Limit scanned data with partitions and clustering,"Use WHERE on partition column; SELECT only needed columns","SELECT * without partition filter","WHERE _PARTITIONDATE BETWEEN '2024-01-01' AND '2024-01-31'"
+BigQuery,Partitioning,Partition tables by date for cost and performance,"PARTITION BY DATE(timestamp_column)","Query without partition filter","CREATE TABLE ... PARTITION BY DATE(created_at)"
+BigQuery,Slots,Understand slot allocation for query performance,"Use INFORMATION_SCHEMA for slot usage; optimize large scans","Ignore slot exhaustion warnings","SELECT * FROM project.INFORMATION_SCHEMA.JOBS_BY_PROJECT"
+BigQuery,UDFs,Use standard SQL before custom JavaScript UDFs,"Built-in functions are optimized","JavaScript UDFs for simple operations","Use SAFE_DIVIDE, IF, CASE instead of JS"
+BigQuery,Approximate,Use approximate functions for large datasets,"APPROX_COUNT_DISTINCT, APPROX_QUANTILES","Exact distinct counts on huge tables","SELECT APPROX_COUNT_DISTINCT(user_id) FROM events"
+BigQuery,Qualify,Use QUALIFY for window function filtering,"QUALIFY ROW_NUMBER() OVER (...) = 1","Subquery wrapper for filtering","SELECT * FROM table QUALIFY RANK() OVER (PARTITION BY id ORDER BY ts DESC) = 1"
+Snowflake,Warehouse,Size warehouse appropriately for workload,"Use X-Small for dev; auto-suspend after 60s","Leave large warehouse running idle","ALTER WAREHOUSE SET WAREHOUSE_SIZE = 'MEDIUM' AUTO_SUSPEND = 60"
+Snowflake,Clustering,Use clustering keys for large tables,"Cluster on frequently filtered columns","Cluster on high-cardinality columns","ALTER TABLE orders CLUSTER BY (region, order_date)"
+Snowflake,Time Travel,Use time travel for data recovery and analysis,"Query historical data: AT (TIMESTAMP => ...)","Forget Time Travel is available for debugging","SELECT * FROM table AT (OFFSET => -3600)"
+Snowflake,Zero Copy Clone,Clone tables for testing without storage cost,"CREATE TABLE test_copy CLONE production","Full physical copies for testing","CREATE TABLE dev.orders CLONE prod.orders"
+MySQL,Indexes,Use composite indexes for multi-column queries,"Create index matching WHERE + ORDER BY columns","Too many indexes slow writes","CREATE INDEX idx_user_date ON orders(user_id, order_date)"
+MySQL,Query Cache,Understand query cache behavior (deprecated in 8.0),"Use application-level caching instead","Rely on MySQL query cache","Use Redis or Memcached for caching"
+MySQL,Limits,Use LIMIT with ORDER BY for pagination,"ORDER BY id LIMIT 100 OFFSET 200","LIMIT without ORDER BY (non-deterministic)","SELECT * FROM users ORDER BY id LIMIT 100 OFFSET 0"
+MySQL,Joins,Prefer JOINs over subqueries when possible,"Rewrite correlated subqueries as JOINs","Correlated subqueries for large datasets","JOIN instead of WHERE x IN (SELECT ...)"
+SQLite,Limitations,Understand SQLite limitations for analytics,"Good for local dev and small datasets","Use for production with concurrent writes","Maximum practical size ~1TB; limited concurrency"
+SQLite,Types,SQLite uses dynamic typing,"Check actual types: typeof(column)","Assume strict typing like other databases","Be aware: '1' and 1 may both be stored"
+Redshift,Distribution,Choose distribution key for join performance,"Distribute on frequently joined column","ALL distribution for large tables","CREATE TABLE ... DISTKEY(user_id)"
+Redshift,Sort Keys,Use sort keys for range queries,"Sort on commonly filtered date columns","Too many sort key columns","CREATE TABLE ... SORTKEY(created_date)"
+Redshift,Vacuum,Run VACUUM and ANALYZE regularly,"Schedule VACUUM DELETE ONLY weekly","Forget to reclaim space after deletes","VACUUM FULL table_name; ANALYZE table_name"
+Redshift,Spectrum,Use Spectrum for querying S3 data directly,"External tables for cold/historical data","Load all data into Redshift","CREATE EXTERNAL TABLE pointing to S3"
+MongoDB,Aggregation,Use aggregation pipeline for analytics,"$match early to reduce documents","Process large result sets in application","db.collection.aggregate([{$match: ...}, {$group: ...}])"
+MongoDB,Indexes,Create indexes for query patterns,"Compound indexes matching query predicates","Index fields not used in queries","db.collection.createIndex({user_id: 1, date: -1})"
+DynamoDB,Queries,Design for query patterns not data model,"Access patterns determine table design","Scan operations on large tables","Use Query with partition key; avoid Scan"
+DynamoDB,GSI,Use Global Secondary Indexes for alternate access,"GSI for different query patterns","Too many GSIs (cost and write amplification)","Create GSI for each major access pattern"
+General,Testing,Test queries on sample data first,"Use LIMIT or sampling for initial development","Run untested queries on full production data","SELECT * FROM table TABLESAMPLE (1 PERCENT)"
+General,Transactions,Use transactions for data integrity,"Wrap related changes in transactions","Auto-commit for multi-statement updates","BEGIN; UPDATE ...; UPDATE ...; COMMIT;"
+General,Comments,Document complex queries with comments,"Add comments explaining business logic","Uncommented complex SQL","-- Calculate 30-day rolling revenue per customer"
+General,Parameterization,Use parameterized queries to prevent SQL injection,"Use ? or :param placeholders","String concatenation for query building","cursor.execute('SELECT * FROM users WHERE id = ?', (user_id,))"

package/assets/shared/data/industries/ecommerce.csv

@@ -0,0 +1,25 @@
+Metric Name,Abbreviation,Category,Formula,Interpretation,Good Benchmark,Related Metrics,Visualization
+Conversion Rate,CVR,Conversion,"Orders / Sessions * 100","Percentage of visits resulting in purchase","2-3% average; 5%+ excellent","AOV, Traffic, Add to Cart Rate",Line chart trend
+Add to Cart Rate,ATC Rate,Conversion,"Add to Carts / Sessions * 100","Percentage adding items to cart","5-10% typical","CVR, Cart Abandonment",Funnel
+Cart Abandonment Rate,Cart Abandon,Conversion,"(Carts Created - Purchases) / Carts Created * 100","Percentage of carts not converted","~70% average; < 60% is good","Checkout Conversion, Payment Failure",Funnel
+Checkout Completion Rate,Checkout,Conversion,"Purchases / Checkout Started * 100","Checkout funnel success","70-80% is good","Cart Abandonment, Payment Methods",Funnel
+Average Order Value,AOV,Revenue,"Total Revenue / Number of Orders","Average amount per order","Varies by category; track trend","Revenue, Items per Order",KPI card trend
+Revenue Per Visitor,RPV,Revenue,"Total Revenue / Total Visitors","Revenue generated per visit","AOV * CVR","CVR, AOV, Traffic",KPI card
+Gross Merchandise Value,GMV,Revenue,"Total value of merchandise sold","Platform transaction volume","Growing MoM/YoY","Revenue, Take Rate",Line chart
+Net Revenue,Net Revenue,Revenue,"GMV - Returns - Discounts - Cancellations","Actual revenue after adjustments","Net/Gross ratio trending up","GMV, Return Rate",Line chart
+Customer Acquisition Cost,CAC,Acquisition,"Marketing Spend / New Customers","Cost to acquire new customer","CAC < first order margin","LTV, ROAS",KPI card
+Customer Lifetime Value,CLV,Acquisition,"AOV * Purchase Frequency * Customer Lifespan","Total expected customer revenue","CLV:CAC > 3:1","CAC, Repeat Rate, AOV",KPI card
+Cost Per Order,CPO,Acquisition,"Marketing Spend / Orders","Marketing cost per order","Track by channel","CAC, ROAS",Bar by channel
+Return on Ad Spend,ROAS,Acquisition,"Revenue from Ads / Ad Spend","Revenue per advertising dollar","3:1+ typically profitable","CAC, CPO",Line chart
+Repeat Purchase Rate,Repeat Rate,Retention,"Customers with 2+ Orders / Total Customers * 100","Customer returning to buy again","> 30% for healthy retention","CLV, Purchase Frequency",Line chart
+Purchase Frequency,Frequency,Retention,"Total Orders / Unique Customers (in period)","Average orders per customer","Varies by category; track trend","Repeat Rate, CLV",KPI card
+Time Between Purchases,TBP,Retention,"Average days between customer orders","Purchase cycle length","Use for remarketing timing","Frequency, Repeat Rate",Histogram
+Reactivation Rate,Reactivation,Retention,"Dormant Customers Who Returned / Total Dormant * 100","Success of win-back campaigns","5-15% typical for campaigns","Churn, Win-back Campaigns",Bar chart
+Return Rate,Return Rate,Fulfillment,"Returned Orders / Total Orders * 100","Percentage of orders returned","< 20% most; < 30% apparel","Net Revenue, COGS",Line chart
+On-Time Delivery Rate,OTD,Fulfillment,"Orders Delivered On Time / Total Shipped * 100","Shipping reliability","> 95% is excellent","Customer Satisfaction",Gauge
+Stock-out Rate,Stockout,Inventory,"Items Out of Stock / Total SKUs * 100","Inventory availability","< 5% for popular items","Inventory Turnover, Lost Sales",Line chart
+Inventory Turnover,Inv Turn,Inventory,"COGS / Average Inventory","How fast inventory sells","Higher is better; varies by cat","DOI, Stockout Rate",Bar chart
+Days of Inventory,DOI,Inventory,"Average Inventory / (COGS / 365)","Days to sell current inventory","30-60 days typical","Inventory Turnover",KPI card
+Website Traffic,Traffic,Traffic,"Total Sessions","Visit volume to site","Growing with quality","Bounce Rate, CVR",Line chart
+Bounce Rate,Bounce,Traffic,"Single-page Sessions / Total Sessions * 100","Visitors leaving immediately","< 40% for product pages","Time on Site, Pages/Session",Line chart
+Pages per Session,PPS,Traffic,"Total Pageviews / Sessions","Engagement depth","2-4 typical; higher for discovery","Bounce Rate, Time on Site",KPI card

package/assets/shared/data/industries/finance.csv

@@ -0,0 +1,24 @@
+Metric Name,Abbreviation,Category,Formula,Interpretation,Good Benchmark,Related Metrics,Visualization
+Gross Profit Margin,GPM,Profitability,"(Revenue - COGS) / Revenue * 100","Profit after direct costs","Varies by industry; 40-60% typical","Net Margin, COGS",KPI card
+Net Profit Margin,NPM,Profitability,"Net Income / Revenue * 100","Bottom line profitability","Positive and stable","GPM, Operating Margin",KPI card
+Operating Margin,Op Margin,Profitability,"Operating Income / Revenue * 100","Core business profitability","Positive for viable business","EBITDA Margin, SG&A",KPI card
+EBITDA Margin,EBITDA,Profitability,"EBITDA / Revenue * 100","Cash-based profitability","Used for comparison across capital structures","Operating Margin, D&A",KPI card
+Return on Assets,ROA,Returns,"Net Income / Total Assets * 100","Asset efficiency","> 5% generally good","ROE, Asset Turnover",KPI card
+Return on Equity,ROE,Returns,"Net Income / Shareholders Equity * 100","Return to shareholders","> 15% considered good","ROA, Leverage",KPI card
+Return on Investment,ROI,Returns,"(Gain - Cost) / Cost * 100","Investment profitability","> 0% means profitable","IRR, Payback Period",KPI card
+Return on Capital Employed,ROCE,Returns,"EBIT / Capital Employed * 100","Efficiency of capital use","> cost of capital","ROIC, Capital Efficiency",KPI card
+Current Ratio,Current,Liquidity,"Current Assets / Current Liabilities","Short-term liquidity","1.5-2.0 typically healthy","Quick Ratio, Working Capital",Gauge
+Quick Ratio,Quick Ratio,Liquidity,"(Current Assets - Inventory) / Current Liabilities","Immediate liquidity","> 1.0 generally healthy","Current Ratio, Cash Ratio",Gauge
+Cash Ratio,Cash Ratio,Liquidity,"Cash / Current Liabilities","Most conservative liquidity","> 0.5 is comfortable","Quick Ratio, Operating Cash",KPI card
+Working Capital,Working Cap,Liquidity,"Current Assets - Current Liabilities","Operating liquidity","Positive and stable","Current Ratio, Cash Conversion",KPI card
+Debt to Equity Ratio,D/E,Leverage,"Total Debt / Shareholders Equity","Financial leverage","< 2.0 generally healthy","Leverage Ratio, Interest Coverage",KPI card
+Debt to Assets Ratio,D/A,Leverage,"Total Debt / Total Assets","Asset leverage","< 0.5 is conservative","D/E, Asset Coverage",KPI card
+Interest Coverage Ratio,ICR,Leverage,"EBIT / Interest Expense","Ability to pay interest","> 3.0 is comfortable","D/E, Debt Service",KPI card
+Asset Turnover,Asset Turn,Efficiency,"Revenue / Average Total Assets","Asset productivity","Higher is more efficient","ROA, Inventory Turnover",KPI card
+Receivables Turnover,AR Turn,Efficiency,"Revenue / Average Accounts Receivable","Collection efficiency","Higher = faster collection","DSO, Cash Conversion",KPI card
+Days Sales Outstanding,DSO,Efficiency,"(Accounts Receivable / Revenue) * 365","Average collection period","Lower is better; industry varies","AR Turnover, Cash Cycle",KPI card
+Cash Conversion Cycle,CCC,Efficiency,"DIO + DSO - DPO","Days to convert inventory to cash","Shorter is better","DSO, DIO, DPO",KPI card
+Revenue Growth Rate,Rev Growth,Growth,"(Current - Prior) / Prior * 100","Revenue increase rate","Depends on stage","YoY, MoM, CAGR",Line chart
+CAGR,CAGR,Growth,"(End Value / Start Value)^(1/Years) - 1","Compound annual growth","Track vs peers and market","Revenue Growth, Projections",KPI card
+Burn Rate,Burn,Cash,"Monthly Operating Expenses - Monthly Revenue","Net cash consumed","Lower = more runway","Runway, Cash Balance",Line chart
+Runway,Runway,Cash,"Cash Balance / Monthly Burn Rate","Months of operations left","> 18 months for fundraising","Burn Rate, Cash Balance",KPI card