Introducing pymyIO — Python bindings for myIO, our d3.js chart library.
34 interactive chart types, composable layer-based API. Renders in Jupyter and in Shiny for Python.
One d3 engine shared with the R package via git submodule — no fork, no drift.
MIT. mortonanalytics/pymyIO
Much of it was due to a migration to reduce data transfer performance. Working with AI and TerraForm, old and duplicated resources arose. Fix: add regular infra monitoring to my dev work; audit everything if anything suspicious shows up.
Quiet day at GroundPulse. No feature, no demo. Raised a timeout, consolidated cloud config, killed long-lived build passwords, added spend alerts. Boring pre-revenue foundation work. Shortcuts you skip now aren't constraints you inherit later.
#InSAR #Geospatial #BuildInPublic
4 months on GroundPulse. Almost none was generative AI.
We built STAC ingestion from NASA OPERA, displacement velocity estimation, seasonal decomposition so alerts don't fire on thermal expansion, z-score thresholds against a multi-year baseline.
Foundation models don't fix broken plumbing.
We built GroundPulse for pipeline displacement monitoring. Then added roadways — same satellite data that flags a moving corridor also flags a failing embankment.
Same physics. Different compliance, different users.
#InSAR #InfrastructureMonitoring #GroundPulse #GovTech
The hard part of satellite infrastructure monitoring isn't the data — it's that each file has 75 million points in a different coordinate system that you have to clip, transform, filter, and match to real assets. Without crashing. That pipeline is the product.
"Does satellite monitoring work in forests?" "What about snow?" "How precise is this?"
If you're evaluating InSAR for infrastructure, these are the right questions. Here's where it actually stands.
Meme about researchers getting raw data
Rust #notajoke
Most operators know where their infrastructure is. Fewer know where the ground underneath it is going.
Satellites measure mm-scale ground displacement every 6-12 days. We built groundpulse.io to turn that into something an integrity engineer can actually use.
#InSAR #InfraMonitoring #GeoHazard
Satellites measure displacement across the full corridor every 6–12 days. The coverage problem is solved at the sensor level.
What's missing: connecting that data to pipeline segments, thresholds, and field workflows. That's what we're building.
Pipeline geohazard monitoring: $300K–$500K/year for a mid-size operator. ILI, ground sensors, aerial patrol, consulting.
Coverage: maybe 5% of the corridor. The known problem sites. The other 400 miles get nothing continuous.
This is also why raw SAR processing isn't the moat anymore. NISAR OPERA products will be analysis-ready — same as Sentinel-1.
The hard part isn't processing radar data. It's connecting displacement to assets, thresholds, and regulatory workflows.
NISAR launched last July, fully operational since January. L-band — penetrates vegetation, maintains coherence in forested terrain. Covers the entire planet every 12 days. Data is already flowing.
We're built to fuse both: C-band where it's strong + L-band where it's strong.
If you manage infrastructure in vegetated terrain, NISAR is the most consequential dataset to come online in years. 🧵
Current InSAR (C-band) decorrelates in forests, crops, wetlands. No. 1 objection from operators: "Your data doesn't work where I need it most."
Fact
Most R charting packages are output-only: data in, picture out.
myIO treats the chart as an input device. Brush to select rows. Click to annotate. Link charts. Slider refits the regression.
Just submitted v1.1.0 to CRAN. 20 chart types, zero extra dependencies.
Satellites have been measuring displacement across every bridge in the country since 2016. Every 6–12 days. A decade of history.
The data that could prioritize a state's $200M repair budget exists. Most DOT bridge programs have never seen it.
#InSAR #IIJA #DisplacementMonitoring
46,154 structurally deficient US bridges gets the headlines. Wrong number.
450K+ bridges rated "fair" on 24-month inspection cycles. Between visits, approaches settle, scour undermines piers, embankments move. All invisible until the next snapshot.
#BridgeInspection #Infrastructure #NBI #InSAR
Map showing moderate to severe soil displacement in the Texas Gulf Coast Corridor
But a colored raster doesn't help an integrity engineer. "Segment 47B needs field validation, confidence 82%, exceeds 3mm/yr threshold" — that's where the actual product lives.
The translation layer is harder than the satellite data.
First time I pulled NASA displacement data over a pipeline corridor, I expected noise.
Got signal instead — clear velocity gradients correlating with known geohazard zones. And movement in places nobody was watching.
We got tired of bolting accessibility onto R charts after the fact, so we built an R/d3.js viz library that starts from 508 compliance and works outward. 20 chart types, SVG, ARIA, keyboard nav. Still in beta — demo at morton-analytics.com/myio/
Client work pays the bills. Product work builds the future. The trick isn't picking one — it's running both so each side makes the other sharper. Slow, deliberate, and the only version of this I'd want to run.
Consolidate config files cutting across over 50 files in an enterprise app makes me - a little nauseous.
Git saves lives - and minds
#debug #rustLang #llm #genAI #codingAssistant #swe #gitSavesLives
But the data doesn't connect to assets, thresholds, or workflows.
The sensor isn't the bottleneck anymore. The translation layer is.
95% of corridor gets zero continuous monitoring. That's not bad engineering — it's bad tooling.
Satellites have been measuring displacement across the US every 6–12 days since 2016. Millimeter precision.
Ground monitoring station: $10K–$100K installed. Watches one point.
Pipeline corridor: 500–2,000 miles. ILI at ~$35K/mile every 5–7 years. Geohazard consulting runs six figures annually.
#InSAR #PipelineIntegrity #InfrastructureMonitoring #GeohazardManagement #gis #rustLang #reactJS #eo
I still thinks it best to bring it up though. Worse case scenario is you are wrong and you learn something
I'm updating my old open source project for the first time in years!
hashtag#rstats hashtag#htmlwidgets hashtag#d3js hashtag#oss hashtag#charts
I love unit testing Math - I can’t really explain why
#math #swe #science #computing #unitTests #rustLang #eo #satellite #radar #SaaS