Methodology
Instruction disruption
Instruction disruption is estimated within the school-day window of 08:00–15:00. When an air alarm overlaps this window, the overlapping minutes are counted as disrupted instruction time. A 5-minute return-to-class allowance is then added after each overlapping event fragment to reflect the time needed to resume learning once the alarm ends.
Homework disruption
Homework disruption is estimated within the assumed homework window of 16:00–20:00. It captures alarm-related disruption outside formal classroom time. No additional return-time allowance is added for this window.
Sleep disruption
Sleep disruption is estimated within the assumed sleep window of 22:00–06:00. It is used as an indirect education-related measure, reflecting conditions that may affect rest, fatigue, and next-day readiness to learn.
School-year filtering
The analytical dataset excludes weekends and assumed vacation periods and retains only agreed school-year windows. The dashboard is therefore intended to describe disruption during periods when learning activity would reasonably be expected to occur.
How events are treated
Alarm intervals that are very close to one another are merged during preprocessing in order to avoid overstating disruption through artificial fragmentation. The dashboard therefore reflects processed disruption events rather than raw source intervals.
About the metrics
Instruction disruption minutes
Total estimated minutes during which teaching time was affected by air alarms within the school-day window. This is the main measure of direct disruption to classroom learning.
Homework disruption minutes
Total estimated minutes during which the assumed homework period was affected by air alarms. This captures interference with learning outside formal school hours.
Sleep disruption minutes
Total estimated minutes during which the assumed sleep window was affected by air alarms. This is an indirect education-related measure reflecting possible effects on rest and next-day learning conditions.
Instruction disruption events
Count of processed alarm event fragments that overlapped the school-day instruction window. This shows how often instruction was interrupted, regardless of how long the interruptions lasted.
Affected instruction days
Number of days on which at least one air alarm disrupted the instruction window. This shows how widely disruption is spread across the calendar, not just how long it lasted.
Students total
Total number of students associated with the selected geography and time context in the education dataset. This is the baseline population for interpreting scale.
Students in person
Number of students studying offline, physically present in schools. This is the most relevant group for direct school-site exposure and shelter-related planning.
Students in person (expanded)
Number of students in offline plus mixed formats. This broader measure captures a wider group of school-connected learners who may still be affected by school-site conditions.
Students online
Number of students studying remotely. These students are less directly exposed to school-site interruption but remain important for contextual interpretation.
Instruction minutes per in-person student
Estimated instruction disruption minutes divided by the number of in-person students. This expresses average direct disruption burden per offline student.
Instruction minutes per in-person student (expanded)
Estimated instruction disruption minutes divided by the expanded in-person population. This provides a broader measure of disruption burden across school-connected learners.
Homework minutes per student
Estimated homework disruption minutes divided by total students. This expresses average out-of-class disruption burden per student.
Sleep minutes per student
Estimated sleep disruption minutes divided by total students. This expresses average night-time disruption burden per student and serves as a proxy for broader educational stress exposure.
Assumptions and interpretation
What this dashboard is for
This dashboard is designed for comparative analysis, prioritization, and public understanding. It helps show where disruption is greater, where students may face heavier burdens, and how patterns change across time and geography.
What this dashboard is not for
The dashboard should not be interpreted as a direct operational record of school-by-school interruption on a given day. It is based on modeled disruption exposure using processed air-alarm data, time-window assumptions, and aggregated education data.
Important interpretation note
All results depend on the defined time windows, preprocessing rules, and available education data. The metrics should therefore be read as analytical estimates rather than exact administrative counts of lost learning time.