A Study on the Improvement of Public Bicycle Services

Focused on South Korean Cases

Yong-Wook Kim and Tea-Seong Lim

Department of Sport Science, Hanyang University, Ansan-si, Republic of Korea

Keywords: Public Health, Public Bicycle Service, Bicycle-Sharing.

Abstract: This study analyzes public bicycle service trends in Korea. First, several cases of small Korean cities with the

highest number of public bicycles were researched. The research results showed that bicycles were used 2,152

times a day with an average of 69 min spent on using the vehicle. Second, the public bicycle service in

Daejeon, a metropolitan in Korea, was analyzed. Factors considered in the utilization rate of bicycles were

neighboring schools, subway stations, and parks. When calculating the expected utilization rate, ten stations

with low usage were assumed to be relocated. In this case, it was confirmed that the utilization rate increased

by 34.00%.

1 INTRODUCTION

Korea introduced its first public bicycle service in

October 2008 after benchmarking the ‘Velib’ system

of France. The service was launched in Changwon, a

regional city with 500,000 residents (S. W. Ha.,

2011). Situated on a flatland with few hills,

Changwon is Korea’s first planned city where people

live relatively close to their workplaces. First, the

local government installed and operated public

bicycle service; however, now a subordinate public

entity is responsible for its maintenance and operati-

on. As the city received credit for its successful public

bicycle service, other local governments began to

introduce such services in their own regions.

Currently, 45 local governments provide 12,453

bicycles for the public through such services. Local

governments can enjoy many benefits when their

citizens use public bicycles often (D. J. Kim, et al.,

2014), which include: fewer cars on the road resulting

in improved traffic flow (C. R. Ye, 2011), significant

reduction in CO2 emissions (H. J. Kwak, K. S. Jung,

2008), healthier citizens due to continual use of bicy-

cles resulting in reduced public health expenditures,

and many more. Comparing the number of patients

visiting hospitals during the period of introducing

public bicycle in each city provides an interesting

result. In Changwon, the rate of patient increase

decreased to 0.67% in 2009 from 1.79% in 2008 when

public bicycle was introduced. In Ansan, the rate

reduced to 0.98% in 2013 from 1.67% in 2012 (Korea

Index website). If you look at the rate of Daejeon

which is the object of this research, it is more reliable.

In Daejeon, the rate was -3.73% in 2012 from 5.92%

in 2011 (as shown in Fig. 1). It’s improbable to

conclude that public bicycles directly reduce the

number of patients in cities. However, it is well

known that regular exercise is good for health

(Centers for Disease Control and Prevention website),

and no one can deny that public bicycle service has

positive effect in regular exercising.

Therefore, policies encouraging more citizens to

use the service should be established. Public bicycles

can replace short-distance transportation so long as

the flow of human traffic is considered when

installing and operating bicycle stations to optimize

utilization. In Korea, public bicycle stations are

mainly located in public sites or areas with large

floating populations (J. Y. Lee, et al., 2012). When

stations attract more users, more bicycle stands are

installed to cope with the increase in usage. In

consideration of the installation of a new station,

citizens’ demand is still a top priority, along with

population flow and the feasibility of using a

particular site. Today, existing public data can be

used to compute an appropriate station location,

including many studies on the improvement of the

functioning public bicycle services (J. T. Wong, C. Y.

Cheng, 2015; S. Wada, et al., 2013; C. Etienne, O.

Latifa, 2014; M. B. Iderlina, 2015; S. Y. Han, et al.,

2013). This allows the calculation of a proper station

location by putting variables together. The purpose of

Kim, Y-W. and Lim, T-S.

A Study on the Improvement of Public Bicycle Services - Focused on South Korean Cases.

DOI: 10.5220/0006058400440050

In Proceedings of the 4th International Congress on Sport Sciences Research and Technology Support (icSPORTS 2016), pages 44-50

ISBN: 978-989-758-205-9

Figure 1: Rate of changes in patients in Daejeon.

this study is to determine where to locate new stations

by using the existing public data and the way to

change existing locations.

2 PUBLIC BICYCLE-SHARING

IN KOREA

After the initial launch in Changwon in 2008, the

public bicycle service is now being operated by 45

local governments with a constant increase in service

adoption. As part of the central government’s bicycle

use promotion initiative launched in 2009, 45,675 km

of cycle tracks have been constructed in South Korea

(Bicycle Happy Sharing website). Today, bicycle

lanes play an active role in providing roads for

Table 1: Public Bicycle Services of Several South Korean

Cities.

City

Service Start Date

Number of Bicycles

In Operation

Daejeon OCT. 2011 2,615

Changwon AUG. 2008 16,611

Ansan SEP. 2012 5,646

commuters in downtowns. Hiking and commuting

courses have been built to connect cities. Bicycle

paths now serve as arteries that allow bicycle users to

travel to most of the cities in the country (as shown in

Fig. 2).

Figure 2: Bikeways of Korea.

A Study on the Improvement of Public Bicycle Services - Focused on South Korean Cases

Table 1 shows local governments that operate more

than 200 public bicycles. In most of the cases, smart

phones are used to rent and return the bicycles.

Financial resources for operations are obtained from

bicycle rental fees, advertising revenues and health

management funds (J. Lee, et al., 2011). Since users

are highly satisfied with the service in most of the

cities, other local governments are considering the

introduction of the service and establishing

infrastructures required for cycle lanes (S. J. Choi,

2011).

3 DATA AND METHODOLOGY

Daejeon, a metropolitan in Korea, was chosen as a

study subject. The city’s public bicycle service called

“Tashu” was analyzed. Data accumulated for three

years from 2013 to 2014 were used for the analysis.

Factors affecting the use of public bicycles were

selected and structural equation modeling was used to

forecast the influences of and demands for the

bicycles. Neighboring schools, subway stations, and

parks, and the population of residences were chosen

as the factors, in consideration of the factors

suggested in preceding research (M. G. Javier, et al.,

2013; N. Gast, et al., 2015; J. T. Pai, S. Y. Pai, 2015;

Y. C. Yoon, B. Y. Cho, 2014). This research

calculates the expected rate of utilization by using

methods below. Factors used for this research are

subway stations, schools and parks. Select 8 local

candidates to find new a' where the rate of Point a

increases. Each point is named a1 to a8. Each local

candidate is located 0, 45, 90, 135, 180, 225, 270 and

315 degrees from north-up direction of points where

Point a to d moves (as shown in Fig. 3). At this point,

Distance d is (station.average)/4 of station distance.

This is due to opinions of public bicycle team of

Daejeon city. The team didn’t want moving too far

from the original stations. Each expected rate of

utilization was calculated from a1 to a8, and factors

used are as follows.

distance.subway(d.su) : Distance to the nearest subway station from Point a

distance.school(d.sc) : Distance to the nearest school from Point a

distance.park(d.pa) : Distance to the nearest park from Point a

influence.subway(i.su) : Influence of subway

influence.school(i.sc) : Influence of school

influence.park(i.pa) : Influence of park

distance.average(d.av) : Average moving distance of users

station.average(s.av) : Average distance between stations

distance.subway1~8 : Distance to the nearest subway station from Point a1 to 8

distance.school1~8 : Distance to the nearest school from Point a1 to 8

distance.park1~8 : Distance to the nearest park from Point a1 to 8

Original.Number(o.nb) : Original Number of Bicycles Rented

New.Number(n.nb) : New Number of Bicycles Rented

((distance.average - distance.subway)influence.subway + (distance.average - distance.school)

influence.school + (distance.average - distance.park) influence.park) : Original.Number = ((distance.average -

distance.subway1) influence.subway + (distance.average - distance.school1) influence.school +

(distance.average - distance.park1) influence.park) : New.Number (f.1)

Therefore,

n.nb1 = (((d.av- d.su) i.su + (d.av- d.sc) i.sc + (d.av- d.pa) i.pa) ((d.av- d.su1) i.su + (d.av- d.sc1) i.sc +

(d.av- d.pa1) i.pa)) / o.nb (f.2)

However, when distance.subway, distance.school, distance.park are bigger than distance.average, use

distance.subway, distance.school, distance.park for distance.average.

icSPORTS 2016 - 4th International Congress on Sport Sciences Research and Technology Support

Figure 3: local candidates of stations.

4 RESULTS

Data related to the use of 202 stations provided by the

Daejeon city were analyzed. On average, public

bicycles were used 2,152 times a day with an average

rental time of 69 min (Daejeon Public Bike website).

The utilization rate during weekends was registered

14.85% while weekdays witnessed an increased rate

of 85.15%. When the daily average was compared for

weekends and weekdays, the daily average during

weekdays was higher with 401 more occurrences of

rental cases. By seasons, spring (April, May and June)

recorded the highest rate with 34.40% whereas winter

(December, January and February) had the lowest

utilization rate of 10.58% (Table 2.). This means that

different weather conditions in the two seasons affect

the frequency of bicycle use (D. J. Kim, et al., 2012).

The time of a day when public bicycles are used most

frequently is 6 PM followed by 8 AM during

weekdays (Table 3.). This suggests that weekday

users use ‘Tashu’ for commuting to work or school,

while weekend users use the service for leisure

activities.

Table 4 shows the analysis of how often users

return public bicycles to stations different from where

they were rented. Bicycle traffic to schools, parks,

and subway stations are as follows: traffic into and

out of schools was 32.14% and 59.02% in the

morning, respectively, while traffic into and out of

subway stations was registered at 28.29% and

24.47%, respectively, in the morning. By contrast, in

the afternoon, bicycle traffic into and out of schools

was 67.86% and 40.98, respectively, and subway

stations saw bicycle traffic flow in of 71.71% and

flow out of 75.53%, respectively. It was confirmed

that public bicycles are mostly used as a means to

commute to work or schools. Public bicycles’ traffic

to and from parks was recorded to be 19.51% and

19.04% in the morning, respectively, while that in the

afternoon was 80.49% and 80.96%, respectively,

showing little difference between inflows and

outflows.

To predict demand, factors affecting demand and the

frequency of use should be compared. The three

factors selected as shown above were set as

independent variables and the frequency of use was

adopted as dependent variables when structural

equation modeling was used for analysis. When it

comes to standardized path coefficients, ‘frequency

of use ← distances to neighboring schools’ was

0.499, ‘frequency of use ← distances to subway

stations’ 0.879, and ‘frequency of use ← distances to

Table 2: Monthly Rental Frequency Average (2014-2015).

Month

JAN. FEB. MAR. APR.

MAY JUN.

Number of Bicycles Rented 23,156 29,460 57,101 80,005 100,877 89,332

Percentage Out of the Year 2.95% 3.75% 7.27% 10.18% 12.84% 11.37%

Month

JUL. AUG. SEP. OCT. NOV. DEC.

Number of Bicycles Rented 87,778 61,071 80,225 88,044 58,013 30,526

Percentage Out of the Year 11.17% 7.77% 10.21% 11.21% 7.38% 3.89%

A Study on the Improvement of Public Bicycle Services - Focused on South Korean Cases

Table 3: Daily Rental Frequency Average by Time (2013-2014, unit: %).

Time 12AM 1AM 2AM 3AM 4AM 5AM 6AM 7AM 8AM 9AM 10AM 11AM

Weekday

Percentage

(per day)

0.80 0.07 0.01 0.00 0.00 0.88 1.52 4.61 8.99 4.55 3.33 3.25

Weekend

Percentage

(per day)

1.18 0.10 0.02 0.01 0.00 0.71 1.07 1.95 3.54 4.18 3.87 4.20

Time 12PM 1PM 2PM 3PM 4PM 5PM 6PM 7PM 8PM 9PM 10PM 11PM

Weekday

Percentage

(per day)

4.02 4.19 4.16 4.61 5.17 6.79 10.85 7.57 6.19 6.60 6.46 5.38

Weekend

Percentage

(per day)

5.43 6.01 6.39 6.70 6.93 7.90 7.97 6.83 6.52 6.53 6.24 5.70

Table 4: Rental Percentage of Bicycles per Major

Area(unit: %).

Area

Percentage of

Bicycles being

Returned

Percentage of

Bicycles being

Rented Out

AM PM AM PM

Stations Located

Near Schools

32.14 67.86 59.02 40.98

Stations Located

Near Subway

Stations

28.29 71.71 24.47 75.53

Stations Located

Near Parks

19.51 80.49 19.04 80.96

parks’ was 0.572. ‘The population of residences’ was

did not affect the frequency of use.

Many studies have proposed algorithms that can

be applied to forecast frequency rates (Y. Seo, et al.,

2015; L. Chen, et al., 2015; Y. Li, et al., 2015; Y. S.

Noh, M. S. Do, 2014). This study, by contrast, adopts

a verified standardized path coefficient to predict the

usage frequency of new stations to be installed. It is

possible to calculate projected frequency rates by

using three independent variables that can affect the

rates. The formula used here is f-2. The same method

was applied to the calculation of the relocation of the

five least used stations. The result is shown in Table

5. It was estimated that the relocation would increase

the overall usage rate by 34.00%.

Table 5: Daily Simulation Results of Bicycles Usage

Frequency Depending on Original and Moved Location

(2013-2015).

Station ID

Original Station

Location

Moved Station

Location

027 28 39

059 32 45

129 58 93

007 63 75

157 69 83

icSPORTS 2016 - 4th International Congress on Sport Sciences Research and Technology Support

5 DISCUSSION

Koreans relatively do more exercising, but when

considering those in their 30s and 40s who are major

participants of economic activities, it is problematic.

As they are busy with their work and due to lack of

time, 60% of them never exercise (Korea Index

website). Bicycle commuting easily provides them

time to exercise. This is the most positive effect of

public bicycle service.

This research removes complex factors and uses

methods to increase the rate of utilization in a simple

way. However, the research to determine station

location is not that simple. There are a lot of

complex factors to be considered. The endless

complexity can be calculated with the development

of computers. Therefore, additional research to

compare the method used in this research with

methods with complex factors used is required.

Simpler method can provide more certain results.

That is why the method to increase the rate of

utilizing public bicycle service confirmed in this

research is very useful. We can search new stations

where the rate of utilization is expected with

analysis considering 3 factors. More accurate

prediction is possible when adding factors including

floating population, income levels, gender and ages.

When measuring the actual utilization rate of

stations with possibility of new installation or

relocation and applying correction value to

algorithm of this research, more effective results can

be provided. Cities planning to introduce public

bicycle service can use methods to determine station

location suggested by this research. They can select

points expected to have high rate of utilization after

first installation and locate stations. positive effect

of public bicycle service.

Mountainous areas account for more than 80%

of Korea, and the favorite exercise for Koreans is

hiking (Ministry of Culture, Sports and Tourism,

2013). The next popular exercise is walking. They

do exercise that they can easily enjoy, and riding

bicycle is also easy. However, there are many slopes

in old cities, and the structure of roads are

inconvinient to ride bicycles. Bicycle roads have

been built in planning stages of new towns, and in

old cities, the roads are mandatory for

redevelopment. The Government is pushing forward

with a national project to connect all cities with

bicycle roads, and bicycle use will continue to

increase in Korea (Ministry of Land, Infrastructure

and Transport website).

6 CONCLUSIONS

This study suggested measures to relocate ten public

bicycle stations currently being operated in the

Daejeon city. The relocation is projected to increase

the frequency from the current 250 to 335. In

addition, if ten more stations are installed, 85

additional occurrences of bicycle use can be

expected, which would increase the frequency by

34.00%. The values were statistically estimated

without considering the operating workforce,

budgets, or influences among stations. Field tests will

be required for correct comparison and verification.

Therefore, this study proposes, first, a preliminary

analysis of new locations for the stations. Second, it

is suggested that the relocation method proposed in

this study be applied to the field to verify the results,

and then a follow-up study be conducted to check if

actual changes are made to the public bicycle service.

We make an inquiry to a public servant involved

in public bicycle service. What are the standards of

selecting stations? The answer was possible areas and

places without complaints which are based on

administrative convenience. As a result, stations are

installed at places with low utilization rate and public

resources are wasted. Basic simulation to determine

installation points is required. In addition, algorithm

to select station location by comparing the actual rate

and expected rate of utilization after operation should

be improved.

ACKNOWLEDGEMENTS

This work was supported by the Research Foundation

of Hanyang University by the Hanyang University of

the Republic of Korea(HY-2016-731).

We would like to thank the anonymous reviewers

for their helpful comments on an earlier version of

this paper.

REFERENCES

Bicycle Happy Sharing. <http://www.bike.go.kr>

(accessed 3.01.16).

Centers for Disease Control and Prevention.

<http://www.cdc.gov> (accessed 3.01.16).

C. Etienne, O. Latifa, 2014. Model-Based Count Series

Clustering for Bike Sharing System Usage Mining: A

Case Study with the Velib System of Paris. ACM

Transactions on Intelligent Systems and Technology.

5(3), Article 39.

A Study on the Improvement of Public Bicycle Services - Focused on South Korean Cases

C. R. Ye, 2011. Public Bicycle Sharing in Washington

D.C.. Monthly KOTI Magazine on Transport. 8, 62-65.

Daejeon Public Bike. <http://www.Tashu.kr> (accessed

3.01.16).

D. J. Kim, H. C. Shin, J. S. Park, H. J. Im, 2012. The Impact

of Weather on Bicycle Usage - Focus on Usage of Bike-

sharing System in Goyang. Journal of Transport

Research. 19(3), 77-88.

D. J. Kim, S. Y. Jeong, S. Y. Han, H. C. Shin, 2014.

Economic Effect Analysis for Bike-Sharing in KOREA

- Focus on Goyang and Changwon City. Journal of the

Korean Society of Road Engineers. 16(1), 63-73.

H. J. Kwak, K. S. Jung, 2008. The Study of Revitalization

for the Public Bike Service -Focused on the "Hi-Seoul

Bike" Project of Seoul. Journal of the Korean Society

of Design Culture. 14(2), 24-33.

J. Lee, S. Kim, Y. Lee, 2011. The management and

operation of Seoul city’s public bicycle business - How

to make profits through the kiosk advertisement.

Journal of Digital Design. 11(1), 231-241.

J. T. Pai, S. Y. Pai, 2015. User Behaviour Analysis of the

Public Bike System in Taipei. International review for

spatial planning and sustainable development. 3(2), 39-

52.

J. T. Wong, C. Y. Cheng, 2015. Exploring Activity Patterns

of The Taipei Public Bikesharing System. Journal of

the Eastern Asia Society for Transportation Studies. 11,

1012-1028.

J. Y. Lee, J. H. Park, Y. T. Leem, 2012. Characteristics and

Strategic Lessons through a Comparative Analysis of

Public Bikes Usage in Three Cities of Korea. Journal

of the Korea Planning Association. 47(3), 295-308.

Korea Index. <http://www.index.go.kr> (accessed 3.01.16).

L. Chen, D. Zhang, G. Pan, X. Ma, D. Yang, K. Kushlev,

W. Zhang, S. Li., 2015. Bike sharing station placement

leveraging heterogeneous urban open data. UBICOMP

'15, SEPTEMBER 7–11, 2015. 571-575.

M. B. Iderlina, 2015. Public Bicycle Sharing in Asian

Cities. Journal of the Eastern Asia Society for

Transportation Studies. 11, 60-74.

M. G. Javier, I. Castillo, A. Queralt, J. F. Sallis, 2013.

Bicycling to university evaluation of a bicycle-sharing

program in Spain. Health Promotion International.

30(2), 350-358.

Ministry of Culture, Sports and Tourism, 2013. A White

Paper on Leisue. Seoul.

Ministry of Land, Infrastructure and Transport.

<http://www.molit.go.kr> (accessed 3.01.16).

N. Gast, G. Massonnet, D. Reijsbergen, 2015. Probabilistic

Forecasts of Bike-Sharing Systems for Journey

Planning. CIKM'15. October 19–23, 2015, 703-712.

S. J. Choi, 2011. A Study on the Design Reformation Plan

of Bike Lane and Storage Facility for Public Space

Design. Journal of the Korean Society of Design

Culture. 17(2), 599-611.

S. Wada, T. Matsuura, K. Numata, 2013. Heuristic Methods

for the Bike-Sharing System Routing Problem. Journal

of Signal Processing. 17(4), 115-118.

S. W. Ha., 2011. Public Bicycle Sharing in ChangWon.

Monthly KOTI Magazine on Transport. 9, 60-69.

S. Y. Han, H. C. Shin, D. J. Kim, 2013. A Comparative

Study for Willingness-to-Pay Estimation Models for

Expansion of Public Bike System. Journal of Transport

Research. 20(2), 103-119.

Y. C. Yoon, B. Y. Cho, 2014. A Study on the Improvement

Plans of Bike-Sharing ‘Tashu’ in Daejeon Metropolitan

City. Journal of Social Science. 25(2), 591-608.

Y. Li, Y. Zheng, H. Zhang, L. Chen, 2015. Traffic

prediction in a bike-sharing system. Sigspatial'15,

November 03-06

Y. S. Noh, M. S. Do, 2014. Analysis of Bikesharing Usage

Patterns and Forecasting Demand Focused on Daejeon

City. Korean Society of Road Engineers. 2014.3, 115-

118.

Y. Seo, D. K. Kim, S. Y. Kho, 2015. A Static Positioning

Strategy for Public Bicycle Systems Considering

Penalty Cost Comparison of Different Demand

Patterns. Journal of the Eastern Asia Society for

Transportation Studies. 11, 1378-1385.

icSPORTS 2016 - 4th International Congress on Sport Sciences Research and Technology Support