Mode

Role: You are a data analyst building a weekly revenue query for Mode. Constraints: produce a single ANSI/Postgres-compatible SQL query that accepts two parameters named start_date and end_date, aggregates revenue by ISO week and product_category, handles timezone by converting event timestamps to UTC, and applies a currency_conversion_rate column when present. Output format: return only the SQL query followed by a 2-line plain-English summary of what each column represents. Example expected columns: week_start, product_category, total_revenue_usd, order_count.

Role: You are a product analyst designing a Mode scheduled PDF report for stakeholders. Constraints: target a single A4 PDF (portrait), include a 1-paragraph executive summary, 4 KPI tiles (metric, comparator, trend arrow), two full-width charts (time series and cohort table), and a one-row notes/footer for data freshness and contact. Output format: return a JSON object with keys: page_size, sections (ordered array with title, visual_placeholder_id, height_percent), metrics (list of KPI definitions), chart_placeholders (ids with chart type and data source reference), schedule (frequency/time). Example: {"page_size":"A4","sections":[...]}

Role: You are a growth analyst preparing cohorts for marketing activation. Constraints: provide a parameterized SQL query that builds acquisition cohorts by acquisition_date using a cohort window of {cohort_window_days} (replace with integer), deduplicates users by user_id, includes device_id, first_acquisition_channel, cohort_start_date, and conversion_date if present, and outputs CSV-ready columns; then include a short Python requests snippet to push the resulting CSV to an HTTP endpoint (show headers and authentication placeholder). Output format: return two labeled sections: "SQL" and "Python (export)" and include one sample row as a comment/example.

Role: You are a data engineer optimizing a slow Mode SQL. Constraints: user will paste their original query between triple backticks below; analyze and return: (1) an optimized SQL query compatible with Postgres/Redshift, (2) 3 short implementation recommendations (indexes, partitions, materialized views), and (3) a one-line estimated percent runtime improvement and why. Output format: JSON with keys original_sql, optimized_sql, recommendations (array), estimated_improvement. Provide explanations in bullets, max 6 bullets total. Paste original SQL here: ```

Role: You are an analytics engineer designing a Mode SQL model versioning and deployment strategy for an enterprise analytics team. Multi-step task: (1) provide a step-by-step workflow (inventory, branching, review, CI, deployment, rollback, audit), (2) include example git branch names, commit message conventions, and PR review checklist, (3) include a sample CI job (YAML) that runs SQL linting, test queries, and schema checks, (4) propose Mode workspace and access-control settings and tagging conventions for model versions. Output format: numbered steps, then code blocks for git examples and CI YAML, then a short FAQ with 3 Q&A about migrations and rollback.

Role: You are a senior data analyst producing an end-to-end funnel analysis in Mode combining SQL and a Python notebook. Multi-step requirements: (1) produce parameterized SQL to compute funnel counts and conversion rates for steps [page_view -> signup -> purchase] by cohort and date range (start_date,end_date); (2) produce a Python notebook outline with named cells for running SQL, loading results into pandas, calculating retention, generating Plotly visualizations (conversion curve, cohort heatmap), and an anomaly detection cell that flags weekly conversion drops >30%; (3) include suggested visualization settings and an alerting rule (condition and Slack payload). Output format: sections titled SQL, Notebook Cells (ordered list with sample code snippets), Visual Config, and Alert Rule (JSON).

Choose Mode over Looker if you prioritize SQL-first notebooks integrated with reporting and code-based analysis workflows.