- Home
- Skills
- Sidetoolco
- Org Charts
- Data Scientist
data-scientist_skill
1
GitHub Stars
1
Bundled Files
2 months ago
Catalog Refreshed
4 months ago
First Indexed
Readme & install
Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.
Installation
Preview and clipboard use veilstrat where the catalogue uses aiagentskills.
npx veilstrat add skill sidetoolco/org-charts --skill data-scientist- SKILL.md976 B
Overview
This skill is a data scientist agent specializing in SQL and BigQuery workflows to deliver fast, actionable data insights. I write efficient queries, run BigQuery (bq) commands when appropriate, and synthesize results into clear findings. I focus on cost-conscious, optimized analysis and practical recommendations. Use this skill to move from question to validated answer quickly.
How this skill works
I start by clarifying the analysis objective, data sources, and constraints. I then design and write optimized SQL for BigQuery, add comments for complex logic, and run queries using bq when needed to manage jobs or export results. I validate assumptions, profile performance and cost, summarize key metrics, and translate results into concise recommendations. Finally, I deliver reproducible steps and next actions.
When to use it
- Exploring large datasets in BigQuery to answer business questions
- Building, optimizing, or reviewing SQL queries for performance and cost
- Generating dashboards-ready aggregates or cohort analyses
- Validating hypotheses with descriptive or diagnostic analytics
- Automating query workflows using bq CLI or scheduled jobs
Best practices
- Clarify the question and required KPIs before writing queries
- Filter early and push predicates to reduce scanned bytes and cost
- Use appropriate aggregations, window functions, and pre-aggregated tables for repeated queries
- Comment complex logic and list assumptions in the analysis notes
- Inspect query plans and slot usage; prefer partitioning and clustering for large tables
Example use cases
- Calculate monthly active users and retention cohorts with a single optimized query
- Compare revenue by channel with join patterns that avoid duplication and overcounting
- Create a cost-aware export pipeline using bq extract for downstream ML training
- Diagnose a slow BigQuery job, identify the bottleneck, and propose partitioning or rewrites
- Produce an executive summary: key metrics, anomalies, and suggested experiments
FAQ
Provide the business question, tables and schemas, time window, desired granularity, and any constraints (cost, latency, or sample size). Permissions and example rows help too.
How do you control BigQuery costs during analysis?
I push filters to reduce scanned data, use partitioned and clustered tables, run LIMIT during iterative development, and recommend scheduled pre-aggregations for production needs.
Will you document assumptions and choices?
Yes. Every analysis includes query comments, a short assumptions list, key findings, and suggested next steps for validation or operationalization.