KSKyle Steinkopf Work with me
Open to data analytics roles & projects

KyleSteinkopf

Data Analyst/AI Developer/Consultant

Turning data into clear reports.dashboards.forecasts.automated pipelines.decisions.

I turn messy data into clear, decision-ready reporting, using Python, statistics and AI to automate the busywork. Backed by years managing data analytics for small and medium business accounts in sales.

Headshot
42 datasetsOne analysis pipeline
Oct 2024Building with AI since
42Datasets in one pipeline
200Daily dials worked in Salesforce
2+ yrsRating AI for TELUS
$30KMonthly sales revenue

01Featured analysis

The Natural Cycles Almanac

An end-to-end data project: 42 public datasets pulled, cleaned and tested in Python, then turned into an interactive dashboard, a plain-language report and a print edition. It's the same work a business needs, from raw exports to answers leadership can act on.

Technical edition · live

From 42 raw datasets to forecasts that are scored, not guessed

  • Python pipeline (pandas, NumPy, SciPy) collects, cleans and validates every source, with automated data and method checks.
  • Every series detrended, then tested: lagged correlations with false-discovery-rate control, wavelet power and coherence, and superposed epoch analysis.
  • Every forecast logged before the season and scored afterwards with the Heidke skill score. Models stay only if they beat a plain trend baseline.
  • Interactive D3 dashboard with eight views, plus a plain-language edition for non-technical readers.
42
Public datasets in one pipeline
8
Interactive views: clock, heatmap, timeline, map, network graph and more
FDR
False-discovery-rate control on every correlation
HSS
Forecasts scored against a baseline, every season
1 · Collect

Pull the data

NOAA weather stations, climate indices, sunspot records, eBird, USA-NPN and Great Lakes ice records.

2 · Clean

Make it trustworthy

Standardize, fill and validate in pandas, detrend so long-term change doesn't pass as a pattern.

3 · Analyze

Test what's real

Correlations, lags, cycles and event studies, with a guard against chance results.

4 · Report

Make it readable

A dashboard for analysts, plain language for everyone else, and a print-ready PDF.

02Services

Clear numbers, less busywork

Reporting, analysis and dashboards, sped up with Python and AI so the routine work takes hours instead of days.

Reporting & Ad Hoc Analysis

Accurate recurring reports and fast answers to one-off questions, from raw exports to a summary leadership can act on.

Dashboards & Visualization

Interactive dashboards and charts that make the trend obvious, like the eight-view D3 dashboard in the almanac.

Data Pipelines & Automation

Python and AI tools that pull, clean and validate data on a schedule, so reports build themselves.

Sales & CRM Analytics

Pipeline, lead and conversion analysis in Salesforce, from someone who managed analytics for small and medium business accounts at Yelp, A Place for Mom and Angi.

Statistics & Forecasting

Trends, seasonality, correlations and forecasts with honest uncertainty, tested against a baseline before anyone relies on them.

AI Development & Consulting

Web apps, internal tools and AI workflows built with Claude Code, plus helping your team adopt AI without the mess.

03Journey

Vibe coding since before it had a name

I started building with AI in October 2024, when the tools needed a person steering nearly every line. That's what I bring to client work: knowing where AI gets it right, where it doesn't, and how to keep it on track.

Oct 2024

First lines with Claude & Grok

Learning to build software by talking it through with the earliest capable coding models.

Prototype

Angelgotchi on Raspberry Pi

The first Angelgotchi, a virtual pet written in MicroPython and running on a Raspberry Pi.

Hardware

ESP32 & C++

Moving to embedded C++ on ESP32 while AI tools still needed a ton of human input and guidance to get anything working.

Full stack

Storefront & backend

Stripe and crypto checkout, a Node/Express API, PostgreSQL order history and Firebase accounts.

3D & games

Worlds in the browser

three.js games, a 3D site map and portfolio, and a 3D changing room wearing a live catalog.

Oct 2026

Data analysis at scale

The Natural Cycles Almanac: a 42-dataset Python pipeline, statistical testing and a D3 dashboard, built with Claude Code.

04Process

Four steps, plain English

You get a clear plan before anything is built, and a team that owns the result after.

  1. 01

    Listen

    A working session on where time and money leak today.

  2. 02

    Map

    A short plan: what to automate, which tools, and what it costs.

  3. 03

    Build

    I build it with AI-assisted development, connect it to your systems and test it on real cases.

  4. 04

    Hand off

    Source code, docs and training so your team owns it, with support while it settles in.

05More work

More built with AI

Everything here was designed and shipped with AI-assisted development. Click through, it's all running.

06Experience

Sales floor to data analysis

Years of managing data analytics for small and medium business accounts in sales at Yelp, A Place for Mom and Angi, so I know which numbers the business actually runs on.

  1. AI Developer, Consultant & Creator

    1333 Co. · founder of 13 <33 CovenantApr 2017 – Present
  2. AI Search Result Rater

    TELUS DigitalApr 2024 – Present
  3. Homecare Acquisitions Manager

    A Place for Mom / AgingCare.comApr 2026 – Aug 2026
  4. AI Research Assistant

    Meta Technologies / Core Casting Group (AI, VR/AR, wearables)Nov 2024 – Nov 2025
  5. Account Executive

    Yelp, Inc.Oct 2023 – Apr 2026
  6. Director of Public Works

    North Irwin BoroughApr 2019 – Nov 2022

Data & analytics

PythonpandasNumPySciPyD3.jsPostgreSQLStatistics & forecasting

Frontier models

Claude & Claude CodeGrokChatGPTMidjourney (since 2024)

Self-hosted

llama.cppLlama 3.1 / 3.2Qwen2.5-CoderQwen2.5-VLBonsai 1-bitKokoro TTS

Image & video

ComfyUIFluxSDXLQwen-Image-EditWan 2.2HunyuanVideoLTX-Video

Development

JavaScriptNode / Expressthree.jsMicroPythonC++ESP32Raspberry Pi

Business stack

SalesforceGoogle WorkspaceStripeFirebasePostgreSQLSlackZoom

Education

B.S., Digital Film Production
The Art Institute of Pittsburgh

07Contact

Need clearer numbers? Let's talk.

Tell me what you need to report on, or what's slowing your team down. I'll tell you honestly what the data can show, and what it would take.

Book a consult Resume PDF