What is a data center?
Wikipedia describes a data center as a physical room, building, or facility used to store, manage, and disseminate data and information, including information-technology infrastructure, computer systems, and associated components.[1] We use this broad category because operators do not always disclose workloads, including whether a facility supports cloud, enterprise, colocation, AI, or other compute.
Coverage
Current coverage includes Houston and nearby regional locations, with expansion to The Woodlands, Brazoria, Galveston, College Station, Conroe, Victoria, and surrounding communities. The map is expanding and is not yet a complete inventory.
Lifecycle labels
Operational or substantially completed facilities.
Evidence says physical construction is underway.
A credible source identifies a planned project, but construction has not been established.
Carrier rooms, network hubs, and smaller colocation sites remain included and are shown by default.
Sources
Our normalized inventory aims to be a digital archive. Data Center Map, Baxtel, operator pages, government records, permits, utilities, and reputable reporting are attached as source records. Directory listings are valuable discovery sources, but their counts and facility boundaries differ.
When a source changes or disappears, we preserve the source URL, access date, captured facts, source role, and available snapshot metadata. A changed source does not silently erase the historical record.
Impact ranges
Beta: The Neighborhood Impact score is still being calibrated and should be treated as a screening aid, not a final assessment. Each detail page displays an estimated range from 0 to 100. The range is not a medical conclusion, property-value forecast, legal opinion, or environmental impact assessment.
Formula: score = regional baseline + sum(category weight × normalized lower/upper input × distance factor), capped at 100. Current weights are Water 25%, Electricity/grid 20%, Air/generation 20%, Sound 15%, Vibration 10%, Land/site context 5%, and Facility scale/site context 5%. Zoning/process, traffic, and wildlife remain contextual evidence in the current model rather than separate direct physical-impact scores. Health is not scored independently because the project does not have exposure or epidemiological data.
Interpretation bands are Minimal (0–19), Low (20–39), Moderate (40–59), High (60–79), and Very high (80–100). A range crossing bands is described using both labels.
Distance and scientific methods
Distance changes each category differently. Water and electricity retain regional floors; air uses a screening proxy informed by the EPA air-dispersion modeling framework; sound and vibration attenuate more quickly using engineering propagation concepts. See the full distance and science guide for formulas and references.
Candidates
Candidates are plausible research records that have not met the verified-record requirements. They are kept in a separate research table and do not appear on the normal home-page map. A candidate can be opened on a map only when it has a responsibly supported location. “Preliminary” is not a separate record type; it describes the uncertainty that makes a record a candidate.
Verified records and unknown data
A verified record has enough evidence to reasonably establish that a distinct facility or project exists, its identity is sufficiently resolved, its location is supported to the stated precision, and its lifecycle label is supported. Verified does not mean every community impact dimension has a facility-specific measurement.
For example, a verified data center may have no public water-use record, sound study, vibration measurement, emissions inventory, waste-heat measurement, or disclosed actual electricity consumption. Those gaps remain visible as unknown, estimated, or proxy values. A verified facility is not a verified prediction of neighborhood health, property value, or environmental impact.
What we exclude
We retain excluded records and their history, but do not place them on the public map when they are power plants without a data-center facility, electrical substations, fiber routes, generic offices or warehouses, industrial land without an identified project, unsupported rumors, or duplicates merged into a better-supported campus record. A cancelled or disproven project is archived or excluded rather than silently deleted.
Electricity scale
When a source lists power in kW or MW, the detailed facility page shows an approximate Texas-home equivalent using 1.5 kW of average continuous household draw. This compares listed capacity, not disclosed actual consumption.
Facilities above 100 homes equivalent are considered significant and start at one electricity-impact point before local distance effects are considered.
Important limits
Unknown data widens a result range and lowers detailed-page evidence confidence. Location precision may be exact, approximate, or candidate. The inventory includes uncertain announced projects intentionally, but users should verify material facts independently before making legal, environmental, appraisal, or inspection decisions.
Use the correction action on a facility page and include the source or public record that should be reviewed.