Slide 1
Slide 1 text
Let's Write a Train Tracking Algorithm
Chris Trott
iOSDC 2025/09/20
Slide 2
Slide 2 text
Hi, I'm Chris
!
From Chicago
"
Now Japan (~8 years)
Slide 3
Slide 3 text
No content
Slide 4
Slide 4 text
Let's Write a Train
Tracking Algorithm
Slide 5
Slide 5 text
What is a train
tracking algorithm?
Slide 6
Slide 6 text
No content
Slide 7
Slide 7 text
No content
Slide 8
Slide 8 text
Talk Overview
1. Review Data Prerequisites
2. Write Algorithm
Slide 9
Slide 9 text
Data Prerequisites
— Static railway data
— Live GPS data from an iPhone on a train
Slide 10
Slide 10 text
Static railway data
— Railways
— Stations
— Railway Directions
— Railway Coordinates
Slide 11
Slide 11 text
Static railway data
— Railways
— Stations
— Railway Directions
— Railway Coordinates
Slide 12
Slide 12 text
No content
Slide 13
Slide 13 text
GPS data
— Local SQLite Database
— sessions & locations tables
Slide 14
Slide 14 text
Location
Slide 15
Slide 15 text
Session
Slide 16
Slide 16 text
No content
Slide 17
Slide 17 text
Write Algorithm
1. Determine Railway
2. Determine Direction
3. Determine Next Station
Slide 18
Slide 18 text
No content
Slide 19
Slide 19 text
No content
Slide 20
Slide 20 text
No content
Slide 21
Slide 21 text
No content
Slide 22
Slide 22 text
No content
Slide 23
Slide 23 text
Example 1
— Railway: Toyoko Line
— Direction: Inbound (to Shibuya)
— Next station: Tsunashima
Slide 24
Slide 24 text
1. Determine Railway
2. Determine Railway Direction
3. Determine Next/Current Station
Slide 25
Slide 25 text
No content
Slide 26
Slide 26 text
Railway Algorithm V1
— Find closest Railway Coordinates to Location
— Sort railways by nearest
Slide 27
Slide 27 text
Railway Distance from Location (m)
Tokyu Toyoko 12.19
Tokyu Shin-Yokohama 177.19
Yokohama Green 1542.94
Tokyu Meguro 2266.07
Slide 28
Slide 28 text
No content
Slide 29
Slide 29 text
However...
Slide 30
Slide 30 text
Example 2
— Railway: Toyoko Line
— Direction: Outbound (to
Yokohama)
— Next station: Hiyoshi
Slide 31
Slide 31 text
Problem: Toyoko Line and Meguro Line run parallel
Slide 32
Slide 32 text
We need history
Slide 33
Slide 33 text
Railway Algorithm V2
— Convert distance to score
— Add scores over time
Slide 34
Slide 34 text
Railway Algorithm V2
Slide 35
Slide 35 text
However...
Slide 36
Slide 36 text
Example 3
— Railway: Keihin-Tohoku Line
— Direction: Northbound
— Next station: Kamata
Slide 37
Slide 37 text
No content
Slide 38
Slide 38 text
No content
Slide 39
Slide 39 text
Equal?!
Slide 40
Slide 40 text
Railway Algorithm V3
— Add penalty for passed stations
— Add penalty for stopping between stations
Slide 41
Slide 41 text
No content
Slide 42
Slide 42 text
No content
Slide 43
Slide 43 text
No content
Slide 44
Slide 44 text
No content
Slide 45
Slide 45 text
No content
Slide 46
Slide 46 text
No content
Slide 47
Slide 47 text
No content
Slide 48
Slide 48 text
No content
Slide 49
Slide 49 text
No content
Slide 50
Slide 50 text
There are many more edge cases but...
Slide 51
Slide 51 text
Let's move on!
Slide 52
Slide 52 text
1. Determine Railway
2. Determine Railway Direction
3. Determine Next/Current Station
Slide 53
Slide 53 text
Every Railway has 2 directions
Slide 54
Slide 54 text
No content
Slide 55
Slide 55 text
No content
Slide 56
Slide 56 text
Example
— Railway: Toyoko Line
— Direction: Inbound (to Shibuya)
— Next station: Tsunashima
Slide 57
Slide 57 text
Direction Algorithm V1
— Mark timestamp for 2 stations
— Compare order of 1st and 2nd station
Slide 58
Slide 58 text
No content
Slide 59
Slide 59 text
No content
Slide 60
Slide 60 text
However...
Slide 61
Slide 61 text
We must visit 2 stations in order to get a prediction...
Slide 62
Slide 62 text
Can we do better?
Slide 63
Slide 63 text
Use Location.course!
Slide 64
Slide 64 text
No content
Slide 65
Slide 65 text
No content
Slide 66
Slide 66 text
No content
Slide 67
Slide 67 text
359.6°
Slide 68
Slide 68 text
Direction Algorithm V2
(1) Fetch 2 closest stations to input
location
Slide 69
Slide 69 text
Direction Algorithm V2
(2) Calculate vector between 2
closest stations for "ascending"
direction
Slide 70
Slide 70 text
Dot Product
Slide 71
Slide 71 text
Direction Algorithm V2
(3) Calculate dot product between
location course vector and closest
stations vector
Slide 72
Slide 72 text
— Positive dot product == "ascending"
— Negative dot product == "descending"
Slide 73
Slide 73 text
-0.95 → negative → "descending" → Inbound
to Shibuya
Slide 74
Slide 74 text
Let's move on!
Slide 75
Slide 75 text
1. Determine Railway
2. Determine Railway Direction
3. Determine Next/Current Station
Slide 76
Slide 76 text
No content
Slide 77
Slide 77 text
— Next: Kawasaki
Slide 78
Slide 78 text
— Next: Kawasaki
— Soon: Kawasaki
Slide 79
Slide 79 text
— Next: Kawasaki
— Soon: Kawasaki
— Now: Kawasaki
Slide 80
Slide 80 text
— Next: Kawasaki
— Soon: Kawasaki
— Now: Kawasaki
— Next: Kamata
Slide 81
Slide 81 text
No content
Slide 82
Slide 82 text
Focus Station Algorithm V1
— Calculate distance d from Location to closest station S
— Calculate direction vector c from Location to S
Case Result
d < 200m "Now: S"
d < 500m && c > 0 "Soon: S"
c > 0 "Next: S"
else "Next: S+1"
Slide 83
Slide 83 text
No content
Slide 84
Slide 84 text
No content
Slide 85
Slide 85 text
No content
Slide 86
Slide 86 text
No content
Slide 87
Slide 87 text
However...
Slide 88
Slide 88 text
GPS data is unreliable
Slide 89
Slide 89 text
Let's create a history for each station
Slide 90
Slide 90 text
struct StationDirectionalLocationHistory {
let stationID: Station.ID
let railDirection: RailDirection
var visitingLocations: [Location] = []
var approachingLocations: [Location] = []
var firstDepartureLocation: Location?
}
Slide 91
Slide 91 text
No content
Slide 92
Slide 92 text
Focus Station Algorithm V2
— Step 1: assign locations to stations
— Step 2: update station phase history
— Step 3: select the focus station phase
Slide 93
Slide 93 text
Step 1: assign locations to
stations
— Assign visitingLocations
or approachingLocations
Slide 94
Slide 94 text
Step 1: assign locations to
stations
— Assign
firstDepartureLocation
Slide 95
Slide 95 text
Step 2: update station phase history
visiting approach departure -> Phase
isEmpty isEmpty !nil departure
isEmpty !isEmpty nil approaching
!isEmpty any nil visiting
!isEmpty any !nil visited
Slide 96
Slide 96 text
approaching
Slide 97
Slide 97 text
visiting
Slide 98
Slide 98 text
visited
Slide 99
Slide 99 text
Step 3: determine focus phase
— Find last station S in travel direction where
phase != nil
Latest Station Phase Focus Phase
departure Next: S+1
approaching Soon: S
visiting Now: S
visited Next: S+1
Slide 100
Slide 100 text
No content
Slide 101
Slide 101 text
No content
Slide 102
Slide 102 text
No content
Slide 103
Slide 103 text
Demo
Slide 104
Slide 104 text
No content
Slide 105
Slide 105 text
github.com/twocentstudios/train-tracker-talk
Slide 106
Slide 106 text
Try Eki Live!
Slide 107
Slide 107 text
✨
Hire Me
✨
— Available for full-time or contract
work
— iOS generalist (not just train
apps)
— twocentstudios.com/blog
— @twocentstudios