Key lithium-ion battery fire statistics
lithium-ion-linked fires in QBE’s returns for 2025.
42 of 49 UK services supplied data [1]e-bike fires in the same QBE collection.
29.5% of its 2025 total [1]e-bike, e-scooter and related-product fire notifications to OPSS.
Voluntary reports for 2025 [2]- QBE’s reported total rose 20.9% in a year, from 1,456 in 2024 to 1,760 in 2025. It was 146.8% above 2022. [1]
- OPSS recorded two fatalities and 98 non-fatal casualties in 2025. These are reported people affected within that collection, not all UK battery-fire casualties. [2]
- Charging was confirmed in 86 of the 267 OPSS notifications. Another 93 were not charging and 88 had unknown charging status. [2]
- London’s published e-bike and e-scooter series reached 206 fires in 2025. The full-year borough tables reconcile to 171 e-bike and 35 e-scooter fires. [3]
- Zurich’s current collection records 172 vape-related fires in 2025, compared with 132 in 2024. [4]
- The widely cited “over 1,200” waste-fire figure is an estimate published in 2024. It was extrapolated from 242 fires reported by 77 councils. [5]
Which battery-fire figures can be compared?
The source label matters as much as the number. A fire involving a product, a notification that a product was the probable cause, and an estimated waste fire describe different measures. The comparisons below keep those definitions intact.
| Source | Latest period | Published measure | Coverage |
|---|---|---|---|
| QBE FOI | 2025 | 1,760 lithium-ion-linked fires | 42 services supplied data; several product types |
| OPSS | 2025 | 267 product-fire notifications | Voluntary notifications; 214 supplied by LFB |
| LFB press / Mayor | 2025 | 206 e-bike and e-scooter fires | London; 33 borough and City areas |
| London Datastore | Jan–Sep 2026 | 414 published incident rows | Broader lithium and electric-vehicle dataset |
| Zurich FOI | 2025 | 172 vape-related fires | 38 responding brigades plus some public data |
| Material Focus | Preceding year, published May 2024 | Over 1,200 estimated waste fires | Survey extrapolation, not a national incident register |
London’s records may appear in both insurer returns and OPSS notifications. Vape and waste fires may also overlap other collections. None of these rows is an additional component to sum into a larger total. [1] [2] [3] [4] [5]
How many lithium-ion battery fires happen each year?
QBE’s May 2026 release reports 713 fires in 2022, 1,138 in 2023, 1,456 in 2024 and 1,760 in 2025. The latest count corresponds to 4.8 recorded incidents a day, or roughly one every five hours, within the responding services. [1]
QBE 2026 FOI collection · responding UK fire services [1]
One publication vintage throughout. Incomplete service coverage and differences in recording affect interpretation.
| Year | Reported fires | Change from previous year |
|---|---|---|
| 2022 | 713 | Baseline |
| 2023 | 1,138 | +59.6% |
| 2024 | 1,456 | +27.9% |
| 2025 | 1,760 | +20.9% |
The increase from 2022 to 2025 is 1,047 recorded fires. It shows growth in the collection, but the release does not provide a complete service-by-year response matrix or a measure of the number of batteries in circulation. It cannot distinguish all changes in incidence from changes in recording, reporting and product ownership.
E-bikes, electric scooters and electric cars
E-bikes are the largest individually named product group in QBE’s published table. Its 520 e-bike fires in 2025 were 39.4% above 2024 and 249.0% above 2022. E-scooter incidents rose from 107 to 182 over the same four-year comparison. [1]
QBE 2025 · product involvement, not fire-risk rates [1]
These groups are within the 1,760 total and do not exhaust it. An involved vehicle is not necessarily the ignition source.
| Product | 2022 | 2023 | 2024 | 2025 | 2022–25 change |
|---|---|---|---|---|---|
| Electric bike | 149 | 272 | 373 | 520 | +249.0% |
| Electric car | 120 | 158 | 207 | 279 | +132.5% |
| Electric scooter | 107 | 134 | 162 | 182 | +70.1% |
E-bikes accounted for 29.5% of QBE’s latest total. The release also reports that 45% involved devices such as phones, tablets, power packs and other battery-powered equipment, but it does not publish a detailed count for each of those devices. This cannot support a phone-versus-laptop league table.
QBE’s updated table and prose both give 520 e-bike fires. An adjacent graphic shows 522 instead; this guide consistently uses the table value and records the discrepancy in the downloadable evidence.
What the official product-safety notifications show
OPSS’s latest workbook records 267 notifications for 2025, compared with a revised 209 in 2024. It includes 217 e-bikes, 46 e-scooters, three mobility scooters and one e-unicycle. This is management information from voluntary reporting, not a national incidence estimate. [2]
| Year | All notified fires | E-bikes | E-scooters | Other products | London notifications |
|---|---|---|---|---|---|
| 2017 | 2 | 2 | 0 | 0 | 2 |
| 2018 | 3 | 2 | 1 | 0 | 3 |
| 2019 | 9 | 7 | 2 | 0 | 8 |
| 2020 | 26 | 10 | 16 | 0 | 26 |
| 2021 | 81 | 51 | 29 | 1 | 80 |
| 2022 | 95 | 75 | 19 | 1 | 87 |
| 2023 | 205 | 167 | 36 | 2 | 176 |
| 2024 | 209 | 166 | 41 | 2 | 174 |
| 2025 | 267 | 217 | 46 | 4 | 214 |
The collection’s relaunch in February 2023 increased participation. In 2025, 18 services supplied at least one notification, and London supplied 214 of the 267 records (80.1%). A rise in notifications therefore cannot be interpreted as an equivalent change in the national probability of a fire.
The current workbook revises earlier figures: 2022 is now 95, 2023 is 205 and 2024 is 209. Older report narratives still display 93, 199 and 211 respectively. We use the latest workbook throughout, rather than joining incompatible editions.
Conversion kits and e-bike build types
OPSS identifies 82 converted e-bikes, 37 manufactured e-bikes and 98 with unknown build type among the 217 e-bike notifications in 2025. Converted bikes represent 37.8% of all notified e-bike fires, while build type is unknown for 45.2%. [2]
OPSS 2025 · 267 notified fires · product/build categories [2]
Unknown build remains visible. Counts lack ownership and usage denominators and do not establish relative product risk.
The 82-to-37 comparison does not establish that a converted bike is 2.2 times as likely to catch fire. That calculation would also need the number of each type in use, how intensively they are used, battery age and consistent reporting. The unknown-build group is larger than either identified build group.
Government guidance nevertheless advises buying a complete e-cycle with compatible battery and charger from a reputable retailer, and warns that poorly fitted or incompatible conversion components can create fire hazards. That advice is supported by product-safety considerations; it is not a risk rate calculated from this table. [6]
Do most battery fires happen while charging?
The OPSS tables do not establish a majority. Charging was confirmed in 86 of 267 notifications (32.2%). There were 93 known non-charging incidents (34.8%) and 88 with unknown status (33.0%). Among the 179 records with a known status, 48.0% were charging. [2]
OPSS 2025 · all 267 notifications [2]
Table-based figures. Time spent charging and not charging is not measured, so these are descriptive shares, not comparative risks.
| Product/build type | Charging | Not charging | Unknown | Total |
|---|---|---|---|---|
| E-bike: conversion | 34 | 30 | 18 | 82 |
| E-bike: manufactured | 11 | 20 | 6 | 37 |
| E-bike: unknown build | 28 | 25 | 45 | 98 |
| E-scooter | 12 | 16 | 18 | 46 |
| E-unicycle | 1 | 0 | 0 | 1 |
| Mobility scooter | 0 | 2 | 1 | 3 |
The published narrative differs by one record: it says 92 not charging and 89 unknown. Both the HTML table and downloadable workbook agree on 93 and 88. We retain those table values without guessing the reason for the discrepancy.
A battery can fail when not plugged in, and missing charging information is substantial. Conversely, the charging count is not proof that a charger caused those fires. The separate ignition-source table records “batteries/generators” in 235 notifications (88.0%), electrical distribution in six, and unknown ignition source in 26.
Where do battery fires start?
OPSS records the property type and the place where a fire began separately. In 2025, 54 notifications were in single-occupancy houses, 44 in purpose-built flats or maisonettes, 13 in converted flats and 11 in houses in multiple occupation. Together these four categories account for 122 notifications (45.7%); that calculation excludes hostels, garages and unknown properties. [2]
OPSS 2025 · location within or around the site [2]
Property type and fire origin are different fields. Other listed locations combines the remaining categories; the full table follows.
| Property type | Notified fires | Share of all 267 |
|---|---|---|
| Street, garden or outdoor area | 66 | 24.7% |
| House single occupancy | 54 | 20.2% |
| Purpose built flat or maisonette | 44 | 16.5% |
| Commercial premises | 16 | 6.0% |
| Converted flat or maisonette | 13 | 4.9% |
| House in multiple occupation | 11 | 4.1% |
| Garage or shed | 5 | 1.9% |
| Hostel or hotel | 4 | 1.5% |
| Road vehicle | 3 | 1.1% |
| Public transport and infrastructure | 2 | 0.7% |
| Unknown | 49 | 18.4% |
View every recorded fire-origin category
| Location where fire started | Notified fires |
|---|---|
| Road or street | 47 |
| Garden or courtyard | 31 |
| Sleeping area | 26 |
| Living area | 24 |
| Corridor or hallway | 23 |
| Garage or shed | 15 |
| Park or open land | 14 |
| Kitchen area | 13 |
| Commercial or office space | 7 |
| Indoor storage area | 6 |
| Balcony or walkway | 4 |
| External structure | 4 |
| Industrial space | 3 |
| Stair or roof space | 2 |
| Road vehicle | 2 |
| Railway or transport area | 1 |
| Bathroom area | 0 |
| Building services area | 0 |
| Outdoor storage area | 0 |
| Unknown | 45 |
Sleeping areas, living areas and corridors or hallways accounted for 26, 24 and 23 notifications respectively. These locations explain why keeping escape routes clear matters, but the data does not measure occupants’ exposure time or establish a risk per home.
QBE separately publishes a wider setting split of 46% in homes, 31% outdoors and 23% in commercial properties. It does not provide the setting-specific denominator or detailed counts, so we retain its percentages without converting them into estimated incident numbers. [1]
Deaths and injuries reported to OPSS
Two fatalities were reported in two OPSS-notified fires in 2025. The collection also records 98 non-fatal casualties across 47 fires. The people and incident counts answer different questions: one fire can affect more than one person. [2]
OPSS notifications · 2020–2025 · people, not incidents [2]
Voluntary reporting and changing participation affect comparisons. Only casualties specifically reported to OPSS are counted.
| Year | Fatalities | Fires with fatalities | Non-fatal casualties | Fires with non-fatal casualties |
|---|---|---|---|---|
| 2017 | 0 | 0 | 0 | 0 |
| 2018 | 0 | 0 | 0 | 0 |
| 2019 | 0 | 0 | 6 | 4 |
| 2020 | 0 | 0 | 18 | 7 |
| 2021 | 0 | 0 | 33 | 16 |
| 2022 | 3 | 3 | 61 | 31 |
| 2023 | 10 | 7 | 78 | 41 |
| 2024 | 7 | 5 | 86 | 43 |
| 2025 | 2 | 2 | 98 | 47 |
Reported non-fatal casualties increased from 86 in 2024 to 98 in 2025, a 14.0% rise. The 2024 fatality total is now seven in the revised workbook, compared with eight in the older report. Missing casualty information does not demonstrate that an incident caused no harm. Fatal and non-fatal casualty fire groups can overlap and are not summed into a count of unique harmful incidents.
London e-bike and e-scooter fire trends
LFB’s published annual series recorded 206 e-bike and e-scooter fires in 2025: 171 e-bikes and 35 e-scooters. That was 20.5% above 2024’s combined 171, following 179 in 2023. The separate Mayor’s full-year borough table independently reconciles to those same 2025 counts. [3]
LFB published annual series · combined incidents [3]
The whole series uses the January 2026 release. Broader Datastore records are presented separately below.
| Year | E-bike fires | E-scooter fires | Combined |
|---|---|---|---|
| 2017 | 5 | 2 | 7 |
| 2018 | 3 | 2 | 5 |
| 2019 | 10 | 2 | 12 |
| 2020 | 13 | 18 | 31 |
| 2021 | 48 | 30 | 78 |
| 2022 | 87 | 29 | 116 |
| 2023 | 143 | 36 | 179 |
| 2024 | 142 | 29 | 171 |
| 2025 | 171 | 35 | 206 |
The LFB release reports two deaths in these fires during 2025 and five since 2023. These London deaths may also appear in OPSS notifications and must not be added to them. The release does not supply a matching annual injury table for this narrower series.
Search battery-related fires by London borough
Lewisham and Southwark each recorded 16 e-bike and e-scooter fires in 2025, followed by Tower Hamlets with 15. These are incident counts from the Mayor’s answer, not a ranking of individual risk or fire-service performance. Borough populations, product ownership and patterns of use differ. [3]
E-bike and e-scooter fires · calendar 2025 · ten of 33 areas [3]
Counts are not adjusted for population, ownership or charging exposure. All 33 areas, including zero-count Sutton, are in the searchable table.
33 areas · 171 e-bike fires · 35 e-scooter fires · 206 combined
| Borough / City | E-bike fires | E-scooter fires | Combined |
|---|---|---|---|
| Barking and Dagenham | 1 | 0 | 1 |
| Barnet | 4 | 4 | 8 |
| Bexley | 3 | 1 | 4 |
| Brent | 5 | 2 | 7 |
| Bromley | 1 | 0 | 1 |
| Camden | 2 | 2 | 4 |
| City of London | 1 | 0 | 1 |
| Croydon | 7 | 1 | 8 |
| Ealing | 4 | 2 | 6 |
| Enfield | 5 | 1 | 6 |
| Greenwich | 8 | 1 | 9 |
| Hackney | 6 | 1 | 7 |
| Hammersmith and Fulham | 5 | 2 | 7 |
| Haringey | 8 | 1 | 9 |
| Harrow | 2 | 0 | 2 |
| Havering | 4 | 0 | 4 |
| Hillingdon | 1 | 0 | 1 |
| Hounslow | 4 | 0 | 4 |
| Islington | 4 | 1 | 5 |
| Kensington and Chelsea | 3 | 1 | 4 |
| Kingston upon Thames | 4 | 0 | 4 |
| Lambeth | 10 | 2 | 12 |
| Lewisham | 13 | 3 | 16 |
| Merton | 5 | 1 | 6 |
| Newham | 5 | 1 | 6 |
| Redbridge | 3 | 2 | 5 |
| Richmond upon Thames | 3 | 0 | 3 |
| Southwark | 12 | 4 | 16 |
| Sutton | 0 | 0 | 0 |
| Tower Hamlets | 15 | 0 | 15 |
| Waltham Forest | 8 | 0 | 8 |
| Wandsworth | 5 | 1 | 6 |
| Westminster | 10 | 1 | 11 |
Sutton’s zero is retained in the complete 2025 table. Blank cells within the Mayor’s published monthly matrices are treated as zero when they reconcile with the published row and grand totals; the original blanks remain documented in the source evidence. No unlisted area has been invented or removed.
| Month | E-bike fires | E-scooter fires | Combined |
|---|---|---|---|
| January | 14 | 4 | 18 |
| February | 12 | 2 | 14 |
| March | 10 | 4 | 14 |
| April | 14 | 2 | 16 |
| May | 14 | 1 | 15 |
| June | 20 | 7 | 27 |
| July | 18 | 4 | 22 |
| August | 15 | 4 | 19 |
| September | 15 | 3 | 18 |
| October | 14 | 3 | 17 |
| November | 10 | 0 | 10 |
| December | 15 | 1 | 16 |
June had the largest combined monthly count, at 27. A single year’s monthly profile is not enough to establish a seasonal pattern.
Latest 2026 London data and why the totals differ
The London Datastore workbook available at the last check reaches September 2026. It contains 414 published rows for January–September, compared with 522 across the whole of 2025. Those periods have different lengths, so this is not a year-on-year increase or decrease. [3]
| Selected workbook category | 2025 full year | Jan–Sep 2026 |
|---|---|---|
| e-bike | 230 | 149 |
| e-scooter | 38 | 28 |
| e-cigarette | 11 | 7 |
| Other Lithium Battery | 93 | 115 |
| Car | 91 | 73 |
| Bus/coach | 28 | 21 |
| All published rows | 522 | 414 |
The workbook includes several vehicle and other-product categories. Its 2025 e-bike count is 230 and e-scooter count 38, while the press/Mayor series is 171 and 35. The sources do not publish an exact reconciliation. We do not assume that every workbook row was caused by battery failure or invent a filter to force the series to match.
| Workbook period | Published fire injuries | Published fatalities |
|---|---|---|
| 2023 | 84 | 4 |
| 2024 | 101 | 1 |
| 2025 | 109 | 3 |
| 2026 (Jan–Sep) | 95 | 4 |
The workbook’s separate casualty sheet reports 109 fire injuries and three fatalities in 2025, then 95 and four in January–September 2026. These are broader-dataset figures. They cannot be attributed to the e-bike rows, individual boroughs or confirmed battery ignition.
The Mayor’s newer borough answer covers only January–August 2026 and reports 109 e-bike fires. Its e-scooter grand total is 19, but the monthly cells sum to 20: Lambeth has one in July and one in August against a printed total of one. Because that conflict is unresolved, the searchable comparison uses the reconciled full-year 2025 table.
Vape and e-cigarette fires
Zurich’s June 2026 analysis lists 172 fires involving vapes, e-cigarettes or their batteries and chargers in 2025. That is 40 more than 2024’s 132, a 30.3% increase. It is a different FOI exercise from QBE’s and should remain a separate series. [4]
UK responding fire services · 2021–2025 [4]
Includes vapes, e-cigarettes, chargers and batteries. Thirty-eight brigades responded; small unreported fires and missing cause data are not captured.
The series rises from 31 in 2021 to 172 in 2025. Zurich’s older October 2024 release used a different sample and period description; its headline cannot be inserted into this sequence. All five values here use the June 2026 edition.
The disposable-vape sales ban began on 1 June 2025, five months into the latest calendar year. Annual incident counts alone cannot establish its effect on fires, especially when product use, disposal behaviour and reporting are also changing.
One example in Zurich’s release concerns a north London newsagent damaged in summer 2024: four fire engines attended, the shop remained closed for more than six months, and the claim cost was about £130,000. This illustrates a possible consequence; it is not an average vape-fire claim.
Battery fires in bin lorries and waste sites
Material Focus’s “over 1,200” UK waste-fire figure is an extrapolated estimate published in May 2024, covering the preceding year. Its December 2025 release describes the period as 2023–24. The March 2026 research still cites that estimate, rather than supplying a new national fire count. [5]
| Step in the estimate | Published figure | What it means |
|---|---|---|
| Councils responding | 77 | Responses to the waste-fire survey |
| Fires they reported | 242 | Observed reports within those responses |
| Average per responding council | 3.14 | 242 divided by 77, rounded |
| Councils used for extrapolation | 382 | The publisher’s UK authority base |
| Estimated UK waste fires | Over 1,200 | Extrapolation, not individually enumerated incidents |
| Councils reporting an increasing challenge | 50 of 53 | 94% of respondents to the detailed question |
The calculation 242 ÷ 77 × 382 is approximately 1,201. The release does not provide an uncertainty interval or a weighting correction for councils that did not respond. Its estimated rise of 71% compares the rounded 1,200 figure with an older estimate of 700 in 2022. Neither estimate should be added to insurer or fire-service totals.
Material Focus’s March 2026 survey adds current disposal context: an estimated 6.3 million vapes and pods were being thrown away or recycled incorrectly each week, compared with 8.2 million in its 2024 comparison. It also found that 47% of surveyed vape purchasers did not know vapes could be recycled and 80% said there was insufficient recycling information.
That disposal research was conducted by Opinium from 29 December 2025 to 9 January 2026 among 1,000 people aged 16 or over who had bought a vape since 1 June 2025. Its survey-based estimates describe purchasing and disposal, not a measured number of fires. Batteries can be crushed or damaged during collection and processing, which is why appropriate electrical and battery recycling routes matter.
Electric-car fires and other electrical incidents
QBE reports 279 incidents involving electric cars in 2025, compared with 120 in 2022. That is a 132.5% increase in recorded involvement. The table does not establish that each fire originated in the traction battery, and it supplies no matched petrol/diesel series or vehicle-use denominator. [1]
A vehicle’s involvement, its fuel or power type, the first item ignited and the recorded ignition source are different fields. This is also why London’s broader vehicle workbook is kept separate from its narrower e-bike and e-scooter publications.
Our vehicle fire statistics and white goods fire statistics cover related topics. Fixed wiring, sockets and mains appliances should not be silently relabelled as lithium-ion battery fires.
What to do about a suspected battery fire
If a battery starts smoking or catches fire, raise the alarm, get out, stay out and call 999. LFB advises against attempting to tackle the fire yourself. Do not carry a smoking or burning battery through the building. [6]
- Use the supplied charger or a replacement authorised by the manufacturer, and follow the charging instructions.
- Keep devices and battery packs away from escape routes and combustible materials.
- Charge only while awake and able to respond; unplug when charging is complete.
- Stop using damaged, swelling or unusually hot batteries and obtain manufacturer or specialist disposal advice.
- Keep batteries out of household rubbish and ordinary mixed recycling. Ask the supplier or council about suitable take-back arrangements.
The statistics show why these precautions matter, but they do not quantify the protection provided by a particular charger, storage cabinet, alarm or extinguisher. For workplaces, assess charging, storage and disposal in the fire risk assessment and communicate the emergency plan. Use staff briefings and fire safety awareness training to explain the alarm and escape arrangements, without encouraging attempts to fight a failing battery.
Frequently asked questions
How many lithium-ion battery fires were reported in 2025?
QBE’s May 2026 FOI release records 1,760 lithium-ion-linked fires for 2025 from services supplying data in that exercise. Forty-two of 49 services responded. It is not a complete official UK census.
How many e-bike fires were recorded?
The QBE table gives 520 in 2025. OPSS separately received 217 e-bike notifications. London’s press and borough series gives 171 e-bike fires, while its broader incident workbook contains 230 e-bike rows. Each figure must retain its source and scope.
Do most reported fires start while charging?
The OPSS tables record charging in 86 of 267 notifications, with 93 not charging and 88 unknown. Even among records with known status, the charging share is 48.0%. These counts do not measure risk per hour of charging.
Can I compare my London borough?
Yes. The searchable table includes all 32 boroughs plus the City of London for calendar 2025, with separate e-bike and e-scooter counts. It describes incidents and is not adjusted for population or product ownership.
Is the estimate of 1,200 waste fires current?
It was published in May 2024 for the preceding year and is still cited by Material Focus’s March 2026 release. It is a council-survey extrapolation, not a new 2026 observed total.
Sources and methodology
Figures are compiled from the original publishers below. Numbered references lead to these source groups; downloadable records retain the exact table or worksheet locator where available.
- QBE: 2026 lithium-ion fire FOI research
Published 11 May 2026. FOI requests sent to 49 UK fire services; 42 supplied data by 21 April. All four years use this release’s updated table.
- Office for Product Safety and Standards: fires in e-bikes and e-scooters
Management information published 30 September 2026. The 2017–2025 workbook supplies revised historical values, product types, charging, property, fire origin and reported casualties.
- London Fire Brigade and Mayor of London: incident data
LFB press release dated 28 January 2026; Mayor’s answer 2026/2755, dated 22 September 2026, supplies borough and monthly tables. The separate Datastore workbook reaches September 2026 and has broader coverage.
Borough tables and January–August update · London Datastore dataset · Original incident workbook
- Zurich: vape fires, June 2026 FOI release
Published 1 June 2026. Thirty-eight brigades responded to requests sent to 47 authorities on 13 April; analysis also uses some publicly available data. All five years come from this release.
- Material Focus: battery fires and waste research
The 10 May 2024 waste-fire estimate is extrapolated from a council survey. It is still cited by the March 2026 vape-disposal research, rather than replaced by a new national fire count.
Latest vape-disposal research, March 2026 · December 2025 clarification of period and survey method
- Department for Transport and London Fire Brigade: battery fire safety
Current official advice checked on 13 November 2026. These sources support the safety section, not an estimate of how many fires happen.
Definitions, revisions and calculation notes
- Source series overlap and have different product, geographical and reporting coverage. No combined UK total is calculated.
- QBE: all years use the May 2026 updated table. Its e-bike bar chart conflicts with the table/prose at 522 versus 520; 520 is retained. QBE setting percentages lack a published detailed denominator and are not converted into counts.
- OPSS: current September 2026 workbook values are used throughout. Older narratives retain earlier counts. The charging table and workbook agree at 86 yes, 93 no and 88 unknown; the narrative differs by one record. No correction beyond choosing the consistent tables is inferred.
- OPSS zero notifications for a listed service do not prove zero actual fires. Unlisted services are not assigned zero. Notification growth is affected by the voluntary scheme and its February 2023 relaunch.
- OPSS property type and location where a fire started are separate classifications. The six named origin categories in the chart total 166; other listed origins total 56 and unknown is 45, reconciling to 267.
- Reported people and fires are distinct units. Fatal fires may also have non-fatal casualties; affected-fire counts are not added. Missing casualty information is not treated as a known absence of harm.
- London 2025 borough data: original complete matrices reconcile to 171 e-bike and 35 e-scooter fires across all 33 areas. Blank matrix cells are zero only within that reconciled table; source values are retained in the evidence.
- London press/Mayor counts and broader Datastore rows are not reconciled by an invented filter. The workbook has no incident identifier or location for its detailed rows; repeated attribute vectors are retained, not deleted as assumed duplicates.
- London partial years differ: Datastore January–September 2026 versus Mayor January–August. The latter e-scooter table has 19 as its printed total but 20 across monthly values; the explorer therefore uses the consistent full 2025 year.
- Zurich: all annual vape figures use the June 2026 edition, not earlier FOI rounds. The 2025 sales ban began partway through the comparison year; no causal effect is calculated.
- Material Focus: the waste-fire estimate comes from 77 council responses, not the separate 4,000-adult recycling survey. The extrapolation has no published uncertainty interval and refers to the period before the May 2024 release.
- Percentages are calculated from the displayed counts unless identified as a publisher’s rounded percentage. Changes use the earlier period as denominator. Counts by product or place do not provide ownership- or exposure-adjusted risks.
- Every chart download retains its source and period. The evidence download includes all OPSS source-table observations, London aggregate records and selected insurer and waste observations; these are overlapping measures, not additive independent facts.
Help your team recognise fire risks
Build practical awareness of prevention, evacuation and workplace responsibilities.
Explore Fire Safety Awareness Training