Structure of Spatial Networks & Communities in LAs Bike Share - - PowerPoint PPT Presentation

Structure of Spatial Networks & Communities in LA’s Bike Share System

Sarah Doggett & Melody Gu CY PLAN 290, May 2018

Purpose

• To evaluate and visualize the Los

Angeles Metro Bike Share using spatial and network analysis

• To identify important stations based
• n use, bicycle flows, and centrality

Zaltz, O’Brien, Strano, & Viana (2013)

• Analysis of bike sharing

systems of five cities:

○ London, Boston, Denver, Minneapolis, and Washington DC

• Similarities and differences
• Considers flows of bicycles

rather than simply using stand

• ccupation data

○ Move beyond Waldo Tolber’s “first law of geography”

• Visualizations of flows (London)

Methods

• Average degree per network to determine station popularity
• Weighted Eigenvector v. Closeness Centrality
• Modularity for Community Detection

○ Case Studies per Community

Data Set

• Trips in Quartile 4 (2017) from October to December, the most recent data

available.

Network Properties

• Number of Nodes/Stations: 120
• Links: 2320
• Users/Trips: 71,214 Trips made in 2017 Quartile 4.
• Average Degree (k) (directed): 75.35
• Shortest Longest Path: 2.42

Community Detection

• Community detection: modularity
• We find 6 main communities:

Community 1 West LA Community 2 East LA Community 3 Port of LA Community 4 North Port of LA Community 5 Pasadena Community 6 Venice

Overall Network Centrality

Community 1: West Downtown LA

TOP 3 STATIONS Degree Centrality Eigenvector Centrality Weighted Eigenvector Centrality 7th & Flower 7th & Flower 7th & Flower 7th & Spring 7th & Spring 7th & Spring Figueroa & 8th Figueroa & 8th 7th & Grand

Community 2: East Downtown LA

TOP 3 STATIONS Degree Centrality Eigenvector Centrality Weighted Eigenvector Centrality Traction & Rose Traction & Rose Union Station W Portal Union Station W Portal Union Station W Portal Main & 1st 1st & Judge John AIso 1st & Judge John AIso Traction & Rose

Downtown LA: Edge Weights

(by community 1 & 2)

Red = 25 or more trips Yellow = 10 - 24 trips Green = 5 - 9 trips Gray = 1 - 4 trips

Community 3: Port of LA

TOP 3 STATIONS Degree Centrality Eigenvector Centrality Weighted Eigenvector Centrality Downtown Harbor Downtown Harbor Cabrillo Beach Fanfare Fountain Fanfare Fountain Downtown Harbor Crafted 22nd St Park Crafted 22nd St Park DoubleTree Hotel

Port of LA: Edge Weights

(by community 3 & 4)

Red = 25 or more trips Yellow = 10 - 24 trips Green = 5 - 9 trips Gray = 1 - 4 trips

Community 4: North Port of LA

Community 5: Pasadena

TOP 3 STATIONS Degree Centrality Eigenvector Centrality Weighted Eigenvector Centrality Old Pasadena Old Pasadena Rose Bowl Pasadena Convention Center Pasadena Convention Center Del Mar Station Fair Oaks & Mercantile Alley Fair Oaks & Mercantile Alley Memorial Park Station

Pasadena Edge Weights

Red = 25 or more trips Yellow = 10 - 24 trips Green = 5 - 9 trips Gray = 1 - 4 trips

Community 6: Venice

TOP 3 STATIONS (All the same!) Degree Centrality Eigenvector Centrality Weighted Eigenvector Centrality Ocean Front Walk & Navy Ocean Front Walk & Navy Ocean Front Walk & Navy Ocean Front Walk & N. Venice Ocean Front Walk & N. Venice Ocean Front Walk & N. Venice Downtown Santa Monica Expo Line Station Downtown Santa Monica Expo Line Station Downtown Santa Monica Expo Line Station

Venice Edge Weights

Red = 25 or more trips Yellow = 10 - 24 trips Green = 5 - 9 trips Gray = 1 - 4 trips

Trip Duration

Average Trip Duration (min.) Community Weekday Weekend 1 West LA 17 29 2 East LA 18 35 3 Port of LA 38 48 4 North Port of LA 27 33 5 Pasadena 19 35 6 Venice 60 64 Los Angeles 25 41

Conclusion

• Prioritize infrastructure improvements to most connected and most traveled

places

• Consider differences in hour of day and weekday/weekend trip behavior

○ This can help with traffic management during high flow periods for both automobiles and bicycles (9AM and 5PM peak) ○ Consider tourism and points of interest for mid-day trips