MESA
Offline geocoding · Spatial cluster detection

Put your cases on the map without putting them on the internet.

MESA is a desktop application for public-health GIS. It geocodes addresses and finds spatial clusters entirely on your own machine. No network, no third-party services, no patient data leaving your computer.

Download MESA What it does →

Windows x64 · macOS Apple Silicon · Linux x64

01 · Capabilities

A field instrument, not a cloud service.

Everything runs locally against your own copy of authoritative address data. No API keys, no rate limits, and nothing to connect to.

Geocode

Addresses, offline

Matches against a local copy of Utah's authoritative address points (UGRC) stored in SQLite, with optional supplemental files for extra coverage.

Analyze

Find the clusters

Local hot- and cold-spot detection with Getis-Ord Gi* and a Bernoulli spatial scan, reported alongside a global Moran's I autocorrelation diagnostic. Monte-Carlo inference, bit-reproducible across every platform.

Prepare

Clean and join tables

Opens Cognos MHT, CSV, and Excel exports. Fix columns, join tables on shared keys, and export to CSV, Excel, or spatial formats.

Protect

PHI stays yours

No telemetry and no outbound connections during analysis. Protected health information never touches a server you don't control.

02 · Workflow

From a spreadsheet of addresses to a map of clusters.

One desktop app carries the whole investigation. Nothing leaves the machine at any step.

Step 01

Load your table

Drag in a Cognos MHT, CSV, or Excel export. MESA parses it, profiles the columns, and tells you what is missing.

.mht · .csv · .xlsx
Step 02

Geocode offline

Map the address column and match every row against a local copy of Utah's authoritative points. Each result carries a precision label.

1,489,417 local points
Step 03

Detect clusters

Run Getis-Ord Gi* or a Bernoulli scan with Monte-Carlo inference to surface the hot and cold spots that matter.

Gi* · Bernoulli scan
Step 04

Export with provenance

Write results back to CSV, Excel, or a spatial format, precision labels and all, ready for the next tool or a report.

.csv · .xlsx · spatial

Every coordinate is labeled, because certainty is not accuracy.

Pointmatched to a real address point, 0 m median error
Interpolatedestimated along a street, block-level
CentroidZIP or city placeholder, locality only
03 · Offline

Built for protected health information.

Spatial epidemiology usually means shipping addresses to a geocoding API. MESA doesn't. The address database lives on your disk, the analysis runs in-process, and the network stays quiet, so you can work on an air-gapped machine and still get reproducible results.

04 · Validation

Validated, not asserted.

The geocoder matches against authoritative data, the statistics are checked against the standard reference implementations, and every run reproduces byte for byte.

1,489,417
authoritative UGRC address points, matched entirely offline
PySAL · R/spdep
Getis-Ord Gi*, Moran's I, and multiple-testing correction cross-checked against the standard libraries
SaTScan 10.3.2
Bernoulli spatial scan matched against the field standard, no discrepancies in the suite
Bit-identical
the same p-values on Windows, macOS, and Linux, from one core or a hundred
Measured accuracy
0 m
median location error for matched address points, with a 95th-percentile error of 28.4 m
~90%
of real SEUHD patient addresses resolve to a rooftop or street point, the rest to labeled centroids
92.7%
sensitivity on the held-out reference test set (99% CI 92.3 to 93.1)
95.7%
precision on the same held-out set, at 98.2% specificity

Measured on an adversarial held-out test partition against positional ground truth, not a synthetic self-benchmark. Cross-validation re-runs in CI against regenerated PySAL, R, and SciPy oracles.

05 · Download

Download

Native builds for each platform. Every release is published with a SHA-256 checksum so you can verify it before running.

PlatformArchitectureStatus
Windows x64 · Windows 10/11 · .exe Coming soon Download
macOS Apple Silicon · arm64 · .dmg Coming soon Download
Linux x64 · AppImage Coming soon Download

Automated per-platform builds are being wired up. When they land, this table fills in automatically from manifest.json with the current version and checksums, with no page to re-edit. Need MESA before then? Get in touch.