Skip to main content

budget-variance-analyzer

Analyze Financial Applications budget-vs-actual variance by cost center and account from aggregate financial ledger tables. Ask me where spend is over or under budget, variance trends, or the biggest variance drivers.

跳到安装

来源信息

仓库
databricks-solutions/ai-governance
最近来源活动
2026年9月2日 15:19
检测到的 SKILL.md 语言
英语
星标
2
分支
2

安装方式

默认使用会先检查来源的 Prompt;你也可以切换为直接命令,或下载本地副本。

检查来源文件

决定是否安装前,请先阅读 SKILL.md,以及 SkillsMP 当前展示的配套文件。

文件资源管理器
2 个文件

正在显示 SKILL.md

SKILL.md
来源说明 · 只读预览
name
budget-variance-analyzer
description
Analyze Financial Applications budget-vs-actual variance by cost center and account from aggregate financial ledger tables. Ask me where spend is over or under budget, variance trends, or the biggest variance drivers.
> **Illustrative example** — demonstrates a well-formed `SKILL.md` for this reference > implementation. Adapt the content to your own org; do not deploy verbatim. # budget-variance-analyzer ## Overview Turns Financial Applications ledger data into **budget-vs-actual variance** insight: it compares actual spend against budget by cost center and account, ranks the largest variances, and tracks variance trends over time. It reads aggregate, department-level financial rollups — no individual transaction PII — keeping it a Tier-2 (internal) finance skill. ## When to use this skill Reach for this skill when a finance user asks about budget adherence: - "Which cost centers are over budget this quarter?" - "What is our budget-vs-actual variance by account?" - "Show me the biggest budget overruns year-to-date." - "How has the IT cost center's variance trended over the last 6 months?" ## Instructions When the user asks a budget or variance question: 1. **Identify the cost center(s), account(s), and period** (default: current fiscal quarter). 2. **Query `greenwood.finance.actuals`** for actual spend by cost center, account, and period. 3. **Join `greenwood.finance.budgets`** on the same keys to compute variance (actual − budget). 4. **Present results** as a ranked variance table, then give 2–3 variance observations (see the Recommendations framework). ## Examples ### Budget variance by cost center (current quarter) ```sql SELECT a.cost_center, SUM(a.actual_amount) AS actual, SUM(b.budget_amount) AS budget, SUM(a.actual_amount) - SUM(b.budget_amount) AS variance FROM greenwood.finance.actuals a JOIN greenwood.finance.budgets b ON a.cost_center = b.cost_center AND a.account = b.account AND a.fiscal_period = b.fiscal_period WHERE a.fiscal_period = DATE_TRUNC('quarter', CURRENT_DATE()) GROUP BY a.cost_center ORDER BY variance DESC ``` ### Variance trend for a cost center (last 6 periods) ```sql SELECT a.fiscal_period, SUM(a.actual_amount) - SUM(b.budget_amount) AS variance FROM greenwood.finance.actuals a JOIN greenwood.finance.budgets b ON a.cost_center = b.cost_center AND a.account = b.account AND a.fiscal_period = b.fiscal_period WHERE a.cost_center = 'IT' AND a.fiscal_period >= ADD_MONTHS(DATE_TRUNC('month', CURRENT_DATE()), -6) GROUP BY a.fiscal_period ORDER BY a.fiscal_period ``` ## Recommendations framework After presenting results, always include: 1. **Biggest overrun** — the cost center/account with the largest positive variance, with the dollar impact. 2. **Trend signal** — is the variance widening or closing over the periods shown? 3. **Quick action** — one corrective observation (e.g. "recurring overrun in account X suggests a budget rebaseline"). ## Edge cases - **Unbudgeted actuals** — if an actual row has no matching budget row, report it as unbudgeted spend rather than treating budget as zero silently. - **Sparse periods** — if the requested period has no ledger rows, say so explicitly. - **Aggregate only** — this skill works at the cost-center/account level. It does not read individual transactions, vendor invoices, or payroll detail. ## Data scope - `greenwood.finance.actuals` — actual spend rolled up by cost center, account, and fiscal period - `greenwood.finance.budgets` — budget amounts on the same keys - No PII or transaction-level detail is accessed — figures are aggregate rollups only.
在 GitHub 查看