What this tool does

It estimates how much of the year a data center can be cooled with outside air alone — free cooling — for any location on Earth, and turns that into energy and money saved. It also shows a two-week outlook so you can see how the coming days compare to normal.

Pick a site, enter a few facts about your facility, and the tool pulls real hourly weather to show where the local climate sits relative to your equipment's allowable operating range. Everything runs in your browser; nothing is uploaded.

The inputs

Location. The data-center site to analyze. Click the map or type latitude and longitude. Hourly temperature and dew point for this exact point drive every result.
IT load [kW]. The average electrical power drawn by your IT equipment (servers, storage, network). Essentially all of it becomes heat that must be removed, so it sets how much cooling energy free cooling can offset.
PUE baseline. Power Usage Effectiveness with mechanical cooling running all year (no free cooling) — total facility power divided by IT power. Typical values are 1.4 to 1.8.
PUE free cooling. PUE while free cooling (economizer mode), when only fans and pumps run and the compressors are off. Typically 1.05 to 1.15.
Equipment classes. ASHRAE TC 9.9 thermal classes A1 to A4. Each defines an allowable temperature and dew-point envelope for IT intake air. Tick the classes you want shown on the map and included in the results.

The yearly density map

The main chart is a two-dimensional histogram of a full year of hourly weather: dry-bulb temperature on the vertical axis, dew point on the horizontal axis. Color shows how many hours the climate spent in each temperature / dew-point cell — brighter means more hours.

The dotted rectangles are the ASHRAE class envelopes (A1 to A4). The more hours that fall inside a class's box, the more of the year that class of equipment could run on free cooling at this site.

The results table

For each selected class the table gives the yearly hours of free cooling (hours inside that class's box), the potential saving in kWh/year, and the potential economy in $US/year.

Savings use a simple, transparent model: the avoided mechanical-cooling power is IT load × (PUE baseline − PUE free cooling), the same facility-wide figure for every class, so only the hours differ. Economy is that energy multiplied by the local commercial electricity price.

This week & next week

Below the yearly view is a short-term outlook built from the current forecast and last year's weather.

Daily strip

The thin strip of bars shows, for each of the next 14 days, how the forecast free-cooling hours compare to what is usual for that date (a yearly baseline). Bars above the line (blue) mean more free cooling than usual; below (red) means less. This week is solid, next week is lighter because the forecast is less certain that far out.

Violin plots

Two split violins compare the distribution of temperature and dew point for this week and next week: last year on one side, the current forecast on the other, with the free-cooling zone shaded. They show at a glance whether the coming weeks are trending cooler (more free cooling) or warmer than a year ago.

Economic tables

For this week and next week, each class's forecast saving (kWh), economy ($US), and the change versus the same week last year.

Data & methodology

Weather (hourly temperature and dew point, historical and forecast) comes from Open-Meteo, based on the ERA5 reanalysis and ECMWF forecasts. The map tiles are OpenStreetMap. Commercial electricity prices are from GlobalPetrolPrices and looked up by country via BigDataCloud reverse geocoding. The equipment envelopes follow the ASHRAE TC 9.9 Thermal Guidelines for Data Processing Environments.

Assumptions to keep in mind. Free-cooling hours are counted as the hours whose temperature and dew point fall inside a class's rectangle — a simplified version of the true psychrometric envelope. The ASHRAE values shipped here are draft figures and should be confirmed against the current standard. Forecasts beyond about a week are less reliable, which is why the next-week view is de-emphasized. Savings are an estimate; tune the IT load and PUE values to your facility.