Data Wrangling
Reliable pipelines, schema validation, and scalable transformations that keep downstream analytics honest.
Available for new engagements
Data Science • Analytics • GIS
Hi, I'm Noah
I turn messy data and raw geography into trusted pipelines, decision-ready analytics, and maps people actually read.
Works across
- R & Python
- Google Earth Engine
- ArcGIS
- Shiny & Quarto
- Mapbox
Key metrics
4+
Years applied experience
25+
Pipelines & apps shipped
100%
Reproducible, documented
End‑to‑end
Ingest, model, visualize
Capabilities
What I do, end‑to‑end
Four pillars that turn noisy, scattered information into decisions your team can defend.
Analytics & Modeling
Reproducible analysis, hypothesis testing, and statistical models that move the KPIs leadership actually watches.
Geospatial & GIS
From TIGER roads to wildfire risk — I build spatial workflows in Earth Engine, ArcGIS, and Mapbox that answer where and why.
Visualization
Dashboards, Shiny apps, Quarto reports, and publication-ready charts that make complex findings feel obvious.
In Practice
Three moves I make on every engagement
Scroll each act to watch the before / after.
raw — before
| id | gender | age | weight |
|---|---|---|---|
| NA | "female" | "twenty-two" | NA |
| 34 | "M" | 34 | 78.2 |
| 999 | "male" | -7 | NA |
| 29 | "female" | 29 | 65.1 |
| NA | "male" | 41 | 9999 |
| 51 | "female" | 51 | 72.4 |
| -1 | "FEMALE" | NA | 61.8 |
| 38 | "male" | 38 | NA |
analysis-ready — after
| id | sex | age | weight | |
|---|---|---|---|---|
| 42 | F | 22 | 68.4 | imputed |
| 34 | M | 34 | 78.2 | |
| — | — | — | — | dropped · outlier |
| 29 | F | 29 | 65.1 | |
| 42 | M | 41 | — | dropped · outlier |
| 51 | F | 51 | 72.4 | |
| — | F | 46 | 61.8 | imputed |
| 38 | M | 38 | 68.4 | imputed |
brittle scripts — before
01_load.R Error in read.csv
setwd("C:/Users/dan/Desktop") · no one else can run this
cleaning FINAL.R object not found
depends on df_raw from a different session · never saved
model_v4_USE THIS.R package missing
library() calls buried mid-script · order matters
make_plots.R never reached
upstream failure · output files missing
reproducible pipeline — after
R/01_ingest.R ✓
here::here() paths · schema validated on load
R/02_clean.R ✓
pure functions · testthat coverage · logged
R/03_model.R ✓
set.seed() · renv.lock · targets DAG
R/04_visualize.R ✓
ggsave() · deterministic · quarto report
ggplot2 defaults
theme_custom()
and so much more.
Selected Work
Things I've built, shipped, and published
Interactive apps, open geospatial datasets, and reproducible analyses — live and in the wild.
Quick Plot
An accessible Shiny app for wrangling, analyzing, and visualizing tabular data — drop in a CSV, ship a chart. (opens in a new tab)
MapLab
Simple GIS environment built for non-GIS users who still need strong map interaction and layer-based exploration. (opens in a new tab)
Map Explorer
Gallery of interactive maps for ecoregions, counties, and other spatial layers designed for exploratory browsing. (opens in a new tab)
US Wildfire Risk
An interactive analysis of wildfire risk in the wildland-urban interface — suppression cost, structural damage, fatalities. (opens in a new tab)
TIGER Roads on GEE
Processed the US Census TIGER road network and published it to the Google Earth Engine community catalog for open, planetary-scale analysis. (opens in a new tab)
WUI Wildfire Explorer
Map-first app for exploring wildfire risk dynamics at the wildland-urban interface across regions and seasons. (opens in a new tab)
Stack
Skills & tools I reach for daily
The technical stack, organized by domain.
Languages
- R
- Python
- SQL
- JavaScript
- Bash
- HTML
- CSS
- Markdown
- LaTeX
Analytics & Stats
- Regression
- GLM / GAM
- Mixed models
- Time series
- Forecasting
- ANOVA / MANOVA
- PCA
- Clustering
- Bootstrapping
- Maximum likelihood
- Hypothesis testing
Geospatial & GIS
- Google Earth Engine
- ArcGIS / ESRI
- Mapbox
- MapLibre
- Leaflet
- OSM
- GDAL
- GeoPandas
- Spatial analysis
- Geoprocessing
Visualization & Apps
- Shiny
- ggplot2
- Quarto
- Jupyter
Data Engineering
- Tidyverse
- NumPy / Pandas
- Git / GitHub
- renv / targets
- Data wrangling
- Pipeline design
- Schema validation
Why Hire Me
What you actually get when you bring me on
Not a consultant who hands you a slide deck — an operator who hands you the pipeline, the repo, and the docs.
Decision-ready narratives
I don't just crunch numbers — I translate them into the two or three sentences an executive needs to make the call.
Reproducible by default
Every analysis lands as a versioned repo: locked dependencies, tested functions, and a README your team can run from a cold start.
Geospatial fluency
Planet-scale raster, national vector, urban analytics — I've shipped against all of them and published open datasets to prove it.
Autonomy, not babysitting
I scope, communicate, unblock myself, and surface tradeoffs early. You get a teammate, not another ticket in your queue.
Let's work
Got a data problem worth solving?
One email. Plain English. I'll tell you whether I'm the right fit — and if I'm not, who is.