About Neyda Larson

I’m a data analyst and AI systems builder in Las Vegas. I turn messy, real-world data into decisions people can act on, and I build the Claude skills and automations that speed that work up.

Portrait of Neyda Larson, data analyst and AI systems builder

What I do

I take a question from raw data to an answer I can defend. I source the data, clean it, model it, build the dashboard, and check every reported number against its source before anyone sees it. Then I build the tools that keep the work running: Claude skills, evaluated against a checklist, and Make.com automations that replace manual steps.

What I've built

Every line below links to a case study you can open and check.

  • SQL, dimensional modeling, Power BI and DAX. A star-schema model and a six-page Power BI dashboard over 143,613 real hospital discharge records. The top 10% of admissions drive 40% of spend, and a few small rural hospitals carry risk the regional average hides. Rural Hospital Risk Hidden by a Healthy Regional Average
  • Python, pandas and statistics. A pre-registered analysis of a federal health survey (7,278 respondents). I caught a variance bug that made the confidence interval about twenty times too narrow, then re-derived every headline number in R as an independent check. The Data Quality Question a Federal Survey Left Unanswered
  • Finding what an average hides. In a B2B sales dataset (900 companies, 2,600 opportunities), the blended win rate looked healthy at 20 to 32% while Enterprise fell from 35.2% to 10.2%. I tested four competing explanations to find the cause. Diagnosing a Segment Failure Hidden by Healthy Aggregates
  • Prompt engineering and conversation design. A Claude conversational agent that replaces a seven-field intake form with a live conversation, plus two coaching skills built on the same ideas. Lead Magnet AI, SQL Instructor, Decision Clarity
  • Evaluated skills. I test my own systems with predictions written before any results exist. On one skill, the case-level average was 100% with the skill and 61.5% without it, three runs per condition. Operator Brain Evaluation
  • Claude agents. A task-tracking skill that keeps its state in one living file, so it survives past any single conversation. Executive Assistant
  • Make.com automation. A five-path lead-qualification workflow that sorts every form submission into a segment and removes manual triage. Make.com Lead Classifier
  • Earlier freelance work with no public write-up. A Flask application that cut a manual reporting process from about 2 hours to 15 seconds (a 99.8% reduction), and a Shopify analytics tool that drove a 12% revenue lift through data-informed product recommendations.

What I'm not

I’m not a software engineer. I have not built production machine-learning systems, run cloud infrastructure, or built vector search. My retrieval work is file-based: an index that links to single-purpose notes. If a role needs those skills, I’m not your person. If it needs someone who finds what the data says, proves it, and explains it plainly, I am.

How I got here

I started in writing and editing. I completed six semesters of biology coursework, then earned a B.A. in Communications from Universidad Central de Venezuela in 2015, and I led a five-person content team at a marketing agency from 2014 to 2016, editing and writing 50 to 100 pieces a day. From 2017 to 2024 I ran Redirect Coaching LLC, a decision-coaching practice, and tracked results across a 35-client cohort: a 46% productivity increase, a 52% anxiety reduction, a 71% sleep-quality improvement, and 83% habit-formation success. In 2024 I moved into data and AI build work.

Credentials

I hold the Google Data Analytics Professional Certificate (2024) and seven Anthropic credentials (2026) covering Claude, agent skills, subagents, and the Claude platform. I also completed LinkedIn Learning courses in Power BI, R, statistics, and project management. See every credential, with verification links where they exist.

Where I'm headed

I’m moving into a formal data analyst role and building toward senior work. I’m based in Las Vegas, NV, and I’m looking for US-based remote roles.

Details: Las Vegas, NV · Remote · English (fluent), Spanish (native)
LinkedIn · GitHub · Let’s talk

Let's Talk