← Back to Blog
fishplate blog

Turning Railway Videos into Searchable Data: How Fishplate Members Can Help

A railway video can contain far more information than its title suggests.

It may feature several locomotives, multiple locations, different operators and one or more reporting numbers. Unless those details are recorded, however, the video can be difficult to find again—especially for somebody searching for a particular locomotive, class, headcode or location.

Fishplate turns public railway videos and their associated metadata into structured, searchable rail records. Instead of leaving each upload as an isolated video, Fishplate connects it to the stock, vehicle classes, operators, locations and headcodes identified within it.

This helps enthusiasts find relevant footage while creating an increasingly useful record of railway video published by hundreds of different channels.

More than 1,500 newly indexed videos

Between 20 June and 19 July 2026, Fishplate added 1,633 video records. Of these, 1,584 were visible in the resulting enthusiast dataset.

Together, those videos produced:

  • 6,951 stock records
  • 2,609 distinct stock identifiers
  • 1,377 headcodes
  • 66 operators
  • 15 vehicle classes
  • 15 indexed locations
  • Videos from 147 different channels

Class 66 produced the largest number of vehicle-class records during the period, with 1,462. Class 37 followed with 555 records, while Class 43 appeared in 331.

Carlisle was the most frequently referenced location, with 236 records. Doncaster appeared 156 times and Lawrence Hill 139 times. Among the reporting numbers, 4L73 appeared in 33 indexed records, followed by 4L45 with 19.

These figures demonstrate the value of structuring railway video metadata. An enthusiast can move beyond searching titles and instead explore videos connected to a particular locomotive, class, operator, location or headcode.

Recent additions include footage whose metadata contains 99006 and 66799 alongside 4M47, a three-locomotive Class 66 combination featuring 66731, 66767 and 66756, and a video of 37254 and 37521 at Brundall.

But building a detailed index at this scale also produces an opportunity: there are records where important information is still missing.

This is where the Fishplate community can help

Railway enthusiasts often recognise details that automated processing or limited source metadata cannot identify confidently.

A locomotive may be visible but not mentioned in the title. A reporting number might be displayed briefly, spoken in the commentary or recorded in the description. The location may be obvious to somebody familiar with the area, even when it has not been named by the uploader.

Adding these missing details makes each video easier for other enthusiasts to discover.

Some current contribution opportunities include:

You do not need to complete every field to make a useful contribution. Confirming one locomotive number, operator, location or headcode can make a video easier to find and improve the connections between Fishplate records.

The important thing is to add information only when it can be identified with reasonable confidence. If a video contains several scenes, locations or vehicles, the details attached to the overall video should not automatically be treated as evidence that every vehicle was filmed at every listed location.

Help make railway video more discoverable

Fishplate becomes more useful as more public videos are indexed—but the knowledge contributed by railway enthusiasts is what helps turn that scale into a more accurate and searchable resource.

If you recognise the stock, operator, headcode or location in one of the incomplete records, take a closer look and add what you can confirm. A small contribution to one video could help another enthusiast find exactly the footage they have been looking for.

Fishplate is based on indexed public railway video and its associated metadata. It is not a complete record of real-world railway activity. Counts and apparent patterns may reflect new uploads, older-video backfill, improved matching, metadata changes or genuine activity. Vehicles and locations associated with the same multi-scene video should not automatically be interpreted as a confirmed sighting of every vehicle at every location.