News and events

Kognitio identifies the Worst Bus Stop in London

The worst bus stop in London has finally been revealed.

Kognitio has used Transport for London’s open data to assess the reliability of TfL’s bus schedules.

Kognitio analyzed 4,948,534,706 data points; 19,687 individual bus stops; 675 bus routes; and 9,641 buses, and can confirm that the worst bus stop in London is:

Ringway, Zone 4
(Towards Southall)

This stop only sees 3.62% of its buses arrive on time, based on the timetabled schedule.

The worst stops in each zone

The worst stops in each zone are:

Zone Stop Name Towards Buses On-Time
Zone 1 Percival Street Angel, Islington 7.57%
Zone 2 Fulham Leisure Centre Putney Heath / Green Man 7.12%
Zone 3 Belthorn Crescent Streatham 3.64%
Zone 4 Ringway Southall 3.62%
Zone 5 New Heston Road Enfield Bus Garage / Southbury Station 3.94%
Zone 6 Quinta Drive Arkley High Barnet 4.6%
Zone 7 Mount Vernon Hospital Pinner, Northwood Hills or Ruislip 9.49%

The best stops in each zone

For comparison, the best performing stops in each zone (which are still quite unreliable) are:

Zone Stop Name Towards Buses On-Time
Zone 1 Clerkenwell Road  Farringdon Rd Blackfriars or Waterloo 27.77%
Zone 2 Stoke Newington Common Hackney Central or Newington Green 22.4%
Zone 3 Ealing Broadway Ruislip Station 27.64%
Zone 4 Wilbury Way Rookwood Road 28.82%
Zone 5 Whitings Road Cockfosters Station 32.83%
Zone 6 Woodlands Grove Wapses Lodge Roundabout 30.68%
Zone 7 Mount Vernon Hospital Rickmansworth, South Oxhey or Uxbridge 13.47%

Further findings, including the most reliable and unreliable routes and post codes, are available here. Kognitio also built a tool which allows users to check how reliable the buses in their post codes are, compared with the rest of London. See the postcode checker at the bottom of this page.

“This project is important” says Roger Gaskell, CEO at Kognitio. “It’s difficult to find interesting and meaningful samples of large data sets so that you can demonstrate big data analytics in action. By pulling data from the TfL API we were able to create a data set large enough to really show what Kognitio can do and to show how, when using Kognitio with a data visualization tool, in this case Tableau, it can run fast, interactive queries against large datasets held in Hadoop . The output is really fascinating as not only does it show the reliability of buses, it has also highlighted how TfL measures its timetables. They measure their performance based on the predicted times on the electronic boards rather than the published “scheduled” timetable.”

“The most interesting part,” continues Gaskell, “is that in some zones the timetable specifies ‘every x to y minutes’ from a given starting time. We wondered how accurate this was. We looked at the waiting times between buses and found that a lot of bus routes don’t arrive at anything like the timetabled intervals. More often than not you do wait ages and then two buses come along together!”