PureWatts AI

    Solar energy assistant

    📍 Home

    Hi there! 👋

    I'm your PureWatts AI assistant. Ask me anything about solar energy, system sizing, or how to use the platform.

    Skip to content

    Solar glossary

    Grid availability: what it means and why it matters

    The share of hours the grid is actually delivering power to a location, derived from night-time lights.

    Why it matters in Nigeria

    It decides how much of a building's energy comes from a generator, and therefore how expensive its energy really is.

    How PureWatts measures it

    Grid avail frac

    n_grid_avail_frac

    Count of buildings in the group. Grid availability, from the NTL proxy.

    Count of buildings in the group.

    Median Grid avail frac

    med_grid_avail_frac

    Median over the buildings in the group. Medians DO NOT COMBINE — never average these across groups; re-derive from the histograms instead. Grid availability, from the NTL proxy.

    Median over the buildings in the group. Medians DO NOT COMBINE — never average these across groups; re-derive from the histograms instead.

    See these numbers for your area

    Every state and local government area page reports grid availability for its own rooftops. Browse the Nigeria rooftop solar data.

    Related terms

    How these numbers are produced

    Approach A — energy value (upper bracket)

    Sizes the system to the building's full latent demand and values that energy at the mixed grid/generator tariff. It answers what the energy is worth — not what anyone currently pays.

    Approach B — cash saved (lower bracket)

    Sizes the system to demand that is actually served today, values it by merit order (solar displaces generator kWh first) and caps it at current spend. It answers what a building stops paying.

    Cohort: Viable rooftops (≥0.6 kWp potential).

    • Demand is modelled latent demand, not metered consumption.
    • Fuel is priced at ₦1,800/litre (operator-set). NBS reported ₦3,277/litre for May 2026 — on that basis every naira figure here is roughly 1.8× understated.
    • Generator ownership is a model output, not a survey result: it follows from the grid-availability scale and the 25% demand-suppression assumption.
    • Approach A prices energy nobody currently buys. Its savings must never be read as cash saved.
    • Demand is MODELLED LATENT demand, not metered consumption.
    • Generator ownership is model output, not survey.
    • Medians do not combine across groups — re-derive from the histograms.
    • Per-building files are float32 with 4 significant figures; cast to float64 before summing, because polars keeps Float32 sums in Float32.
    • Null scenario fields mean the roof cannot host 0.6 kWp, not zero savings.
    • `genset` was previously named `diesel` and now uses the segment's real fuel.
    • Ordering holds by construction: grid savings < mixed < genset.

    Frequently asked questions