Designing and Evaluating Iter: A human-centered carpooling
application for students of the Costa Rica Institute of Technology
SUBMITTED IN PARTIAL FULLFILLMENT FOR THE DEGREE OF MASTER OF SCIENCE
Andres Elizondo
11104376
M
ASTER
I
NFORMATION
S
TUDIES
H
UMAN-
C
ENTERED
M
ULTIMEDIA
F
ACULTY OF
S
CIENCE
U
NIVERSITY OF
A
MSTERDAM
July 21, 2016
1st Supervisor 2nd Supervisor Dr. Cesar Garita Dr. Frank Nack SCC, Costa Rica Institute of Technology ISLA, UvA
Designing and Evaluating Iter: A human-centered carpooling
application for students of the Costa Rica Institute of Technology
Andres Elizondo
University of Amsterdam, Faculty of Science. Science Park 904
1098 XH, Amsterdam andres.elizondorojas@student.uva.nl
ABSTRACT
This article analyses in depth the issues affecting mobility in Costa Rica; ranging from a broad set of socioeconomic, governmental, environmental and technological factors. These elements served as a basis to design Iter: a carpooling application prototype that was tested with students of the Costa Rica Institute of Technology in order to encourage vehicle sharing, and to analyze with a human centered approach what are the possible benefits or issues while using these type of services. In order to accomplish this, the investigation was divided in three main phases: An in-depth theoretical research to provide a clear problem framework, and to extract the requirements necessary to effectively develop the platform. Then, the paper focuses thoroughly in explaining the modular architecture used and how the system was finally designed using the most cost-effective, user-centered and agile components found during the research. Lastly, the investigation uses a qualitative approach to analyze from the users’ perspective the impact of using a vehicle sharing application. The results of this investigation try to set a solid foundation on the perceived benefits in the mobility of a region by using this type of sharing-economy platforms. Moreover, the system was designed in such way that other institutions and user groups can be incorporated to the system in the short-term, focusing on Latin America as the main target region for future developments and scalability of the application.
Keywords
Human centered multimedia; transportation; carpooling; sharing economies; mobile systems; interaction design.
ACM Classification Keywords.
H5.2. Information interfaces and presentation (e.g., HCI): User Interfaces (User-centered design).
1. INTRODUCTION
Mobility has a great impact on the life quality of people, and it directly affects the whole environment of a city. From pedestrians, passengers and those who commute in their private vehicles, or using public transportation, all of them are active players of the mobility of a city. In 2013 more than 121 million people living in the United States commuted daily to their jobs or their study institutions [1]. Of those people, more than 90% use their own private vehicles to commute, but approximately only 9.7% of them share it with colleagues or other students.
Studies showed that just by shifting to a vehicle sharing culture – decreasing the number of journeys and increasing the density of passengers per vehicle–, each year, CO2 emissions can be lowered from 8.96 to up to 29.97 tons, and about 336 million liters of fuel can be saved annually [2][3][4]. In other aspects, carpooling alleviates a big part of the problems related to high traffic in cities, availability of parking spaces, time necessary to commute and it even generates a greater fellowship among coworkers and students.
These are all examples of how it is necessary to promote in all levels of society (government, institutions, companies and individuals) a new cultural behavior to influence mobility and improve the way people commute. For this reason, this research was focused on leveraging the possible benefits of a carpooling community, to accomplish this, a mobile application prototype was developed and tested with students of the Costa Rica Institute of Technology (TEC). In addition to measuring the acceptability of the proposed solution, the research analyzed the effort put into similar projects and adapted that work to the characteristics of the Latin-American, highly-educated and youthful population –often referred to as millennials- used as the research target group. Besides the technical aspects of the prototype, an intensive research on the concept of sharing economies was performed. Conceptually, this model directly influences the way products and services are obtained and used within a society; this can be done by exchanging, renting or sharing goods with other people, instead of buying them from companies as it occurs in a classical capitalism. Essentially, sharing economies allow the society to make a more efficient use of products they own. This efficiency also applies indirectly to other aspects, for instance, better use of resources (gas, water, and many others), and helps to minimize environmental problems such as the pollution generated by vehicles. This way, providers create a better and more diverse supply of services and products, and also, although not necessarily, they can make –even a big– profit with it. On the other hand, consumers are greatly benefited from the model as they usually find better conditions for what they are looking for: lower prices, better services, and more convenient locations, just to mention a few.
The importance of having a strong work theoretical framework in the concepts of carpooling and sharing economies, lie in the sociocultural, economical, governmental and environmental characteristics of the costarrican region that this research pretends to encompass as future target group. The Great Metropolitan Area of Costa Rica (GAM for its acronym in Spanish) is the most urbanized, populated and economically active region of the country, and represents the administrative, political and cultural center of Costa Rica. According to the National Institute of Statistics and Census (INEC), approximately 52.7% of the population resides in this area of just 2044 km2 (about 3.8% of the territory) and it is composed by 31 cantons that belong to four of the seven costarrican provinces [5]. Moreover, it comprises more than half of the active working population of the country; from which about 75% work on the services economical sector, 21% in the industry and the remaining 4% in agriculture. Of all workers from the GAM, 49.4% work in the same canton that they live, 46.1% in other cantons within the GAM, and only 4.5% work outside the GAM or outside the country. This figures summarizes in a relatively small but densely active urban area, with a significant number of inhabitants living and commuting daily on it.
2. RELATED WORK
In order to conduct the research, a review of the current state-of-art in mobility applications and the concepts behind sharing economies was performed. The objective was to gain as much insight as possible in these subjects; first, to adapt this knowledge to the costarrican culture in which the prototype was tested, and after the research was done, being able to escalate the project to other groups and countries in Latin America.
2.1. Carpooling
During the investigation, the first study approach used was understanding and defining which carpooling applications existed on the market, and which studies on the topic were already performed. This was intended to extract their main characteristics, both in positive and negative aspects. Furthermore, this allowed to have a valuable framework to start with the system design, which was complemented by the subsequent phases of the research. First, the social and environmental benefits of carpooling were investigated. Initially, the National Report on Commuting Patterns and Trends from the American Association of State Highway and Transportation Officials described many of the relationships between the socio-demographic characteristics of commuting people, the geographic distribution of residence and work places, the quality and availability of various modal options and commuting mode choices of these commuters [1]. This was a good initiation for the study as proved that although several carpooling services already exist in the United States, people still continue traveling alone or not sharing their vehicle with other passengers. As it is going to be explained in the Context Statement of this research, this was helpful to investigate if the physiognomy of the citizens of Costa Rica matched these characteristics, helping to understand what social features needed to be addressed first. Then, the potential advantages of carpooling were analyzed. In this part of the research, three investigations were of utterly importance as they highlighted and provided evidence of the environmental benefits of using this type of systems. The investigation of Farrell et al. described the potential success of promoting and encouraging the use of soft modes of transport; shifting conventional transportation modes to walking, cycling, and carpooling [2]. The research focused on the possible reductions of CO2 emissions to the environment, and concluded that no standalone solution works as a unique approach to achieve substantial benefits for the environment, instead, a mix of all these options is necessary in order to fulfill that objective. Caulfield then explains in his research that carpooling not only alleviates a great part of contaminating CO2 emissions, but the study performed in Dublin showed that the type of occupation and age of the citizens directly influence the decision of carpooling or not [3]. Caulfield states that skilled and non-skilled people, such as students, and those working in manual industries, as well as younger populations, are more likely to ride-share. Lastly, the fuel saving aspects of carpooling were described in the paper by Seyedabrishami et al., the study observed the tendencies of carpooling travelers in the city of Tehran, capital of Iran [4]. It used a mathematical analysis employing a survey among its participants and concluded that about 780000 vehicle trips could be saved per day, representing more than 336 million liters of fuel annually. Moreover, it concluded that just by using websites or technologically-enabled practices, the number of vehicle sharing individuals could increase up to 30%, resulting in even greater fuel savings and less contaminating emissions released to the environment.
Many carpooling applications are also available on the most popular app markets, but most of them are available or only being used in certain regions or countries. Specifically, this research found out that applications and platforms such as Blablacar in Europe, and Coride and Rideshare in the US, are examples of successful companies focused on carpooling that have great acceptance in those regions. In the other hand, although projects such as PEMS in Colombia and Workride in Costa Rica try to provide similar services, the efforts put into this specific region seem to not have a widespread introduction of a vehicle sharing culture in Latin America yet. Therefore, part of the objectives of this study is to design a system that encourages active participation of users, availing the features and requirements extracted from the Context Statement of this paper.
2.2. Sharing Economies
A fundamental part of this investigation was to provide strong theoretical framework of sharing economies. Both in terms of what the concept means, and also providing examples of how this type of economy can benefit and change the way goods and services are acquired, offered, and shared between people.
Farid Tabarki states in his book “Het einde van het midden” (The end of the middle), that shared economies are dissolving the need to have a “middle-man” between users, therefore shifting the concept of capitalism and how people and companies operate within society [6]. This creates a new ecosystem in which providers and consumers –prosumers as named by Rifkin – are essentially at the same level [7].
Recently, the model proposed by sharing economies is rapidly being introduced into our society, platforms like Airbnb (home sharing app), Uber (taxi-like service) and Prestadero (Mexican service for loans) are current, modern and popular examples of this [8][9][10]. The exponential growth of the internet, social networks, and mobile devices and applications have contributed enormously to this type of peer-to-peer platforms. This whole infrastructure environment allows the cost of each transaction to be lower, and facilitates to replicate the model in large scale as stated by Rifkin. Technologies such as geolocation, online payments, and mobile systems are allowing people to find others providing the service or product they are looking for in an efficient and easy manner. In theory, this is something that people were able to do before the Internet and mobile devices, but was much more difficult to do; so much, that people preferred buying a product, or paying to a company to provide them a service, even when the cost was higher. The model of a sharing economy works even better with products or services that are abundant, expensive to obtain, and not being used all the time, the most common example of this are houses and cars; it is no surprise that Airbnb and Uber took over those markets long time ago. Also, there are many other platforms currently emerging that allow people to share other spaces such as parking spots, camping areas and even appliances like laundry machines, as it is being done in France by La Machine du Voisin [11].
Furthermore, social networks are helping to promote trust among providers and consumers as the information provided by these platforms allow users to verify the identity of both individuals [12]. They also let to check any social connections that they might have in common, and even to crowdsource information about previous interactions with other users [13]. This generates a whole trustworthy environment in which users gently feel safer and become more confident.
Nevertheless, despite all the benefits mentioned before, our society is being through a great regulative and sociocultural change; it is not yet fully adapted to this kind of new economy, as mentioned by Tabarki and Rifkin. Services like Uber and Airbnb not only remain illegal in many countries, but they are currently going through many legal trials where they are accused of providing illegal services, not having the same insurance coverage as taxi services or hotels, and they are even suspects of alleged tax avoidance. Furthermore, although many of these platforms provide ways to have insurance, or make claims if there are issues, the main responsibility belongs to the users themselves. As it was stated before, in this model the prosumers are situated at the same level, so typically there is no authority or figure that takes all the responsibility if a problem arises.
Even more, Uber and Airbnb are not the best description of a fully shared economy, as they are still companies that profit with the transactions done by prosumers. However, the positive aspect is that they are stimulating a new culture of sharing among people, and helping them to realize that there is the possibility to choose the services or products that fulfill their demands best. This part was crucial during the investigation, as the main idea behind the research is promoting an operational vehicle sharing community in which there is no company profiting from the services provided. For all the reasons mentioned before, it was of utterly importance for the research to analyze the implications and any possible issues that could arise from promoting a carpooling community. To go even further, future research will help to analyze if the introduction of a decentralized architecture and applications (for example, blockchain technologies such as Ethereum [14]) can provide a positive effect for this specific type of communities and services.
3. THESIS STATEMENT
The following section defines the main research question and the objectives that this study foresees to achieve. This section is closely linked to Methodology, which explains how these goals are planned to be accomplished throughout the research.
Research Question
How should the application prototype be designed using a human-centered approach in order to encourage vehicle sharing
among students of the Costa Rica Institute of Technology?
Research Objectives
Identify the socioeconomic, governmental, environmental
and technological aspects that influence mobility in citizens of the GAM of Costa Rica and TEC students.
Define the system components that allow to develop the
application prototype in the most cost-effective, user-centered and agile way.
Evaluate the proposed prototype design and obtain the users’
perception about carpooling platforms.
Expose any findings relevant for future scalability of the
prototype to other groups and countries in Latin America.
4. METHODOLOGY
In order to conduct this research, a mixed-methods approach was used, consisting of three main phases. First, using an in-depth theoretical analysis approach, the problem context was investigated in order to collect sufficient information to define the main research question and objectives, and also to construct a strong framework to start with the prototyping phase. The next method then consisted of an engineering research approach in order to develop and test the application prototype. Lastly, a qualitative research was performed in order to obtain the results of the investigation, answer the main research question and verify if the study objectives were achieved.
4.1. Theoretical Research
This phase took place in the early stages of the investigation, and consisted in an in-depth literature study and framework examination; exploring the sociocultural, economic, environmental and technological context around mobility in Costa Rica and the feasibility of developing a carpooling application for TEC students, and eventually for the GAM region.
The analysis was allowed by the definition of the research question and objectives: how to effectively design the application prototype, and what characteristics had to be taken into consideration for the system design. This information was crucial for the whole investigation, and was the base from which the following phases of the research were implemented.
4.2. Prototype Development
In this phase, the application prototype was developed and then tested with TEC students. The aim of this engineering research method was to find the most cost-effective, user-centered and agile components to build the system in the short time given for this research.
The prototype was developed based on the requirements and features extracted from the theoretical research. The main objective of this phase was to provide a system design that could not only fulfill the objectives of the investigation, but also serve as a model for other projects and applications given its modular architecture and its market-proven principles used to design the whole platform. This part of the research was also important as it served as a user information database used in the Qualitative Research. When testing the application, users provided their contact data which was later used to inquire and encourage them to participate in the qualitative study explained next.
4.3. Qualitative Research
The final phase of the investigation consisted in obtaining the qualitative data extracted from a semi-structured interview approach. Essentially, in this part of the research, the perception of the students that tested the application prototype was acquired and analyzed. The primarily objective was to extract results and formulate conclusions out of the people’s perspectives, focusing on two main areas: their thoughts about the overall design and the interactions developed for the application prototype, and a subjective survey about carpooling platforms, and the benefits and problems they feel that these type of platforms carry with them. After the information was obtained, the investigation focused then on matching these results to the main research question and objectives. Lastly, the outcome of this stage served as a basis to formulate the results and conclusions of the investigation.
5. PROBLEM STATEMENT
This section thoroughly explores the problem context behind the research. It helped not only to fully understand the problems affecting mobility in Costa Rica, but it fully details each major factor taken into consideration for the development of the application prototype. It is intended to provide a great socioeconomic, governmental, environmental and technological framework to the reader, and it provided abundant knowledge to achieve the objectives of the research.
5.1. Mobility in Costa Rica
Even when the GAM provides a decent quality of life to its citizens, there are many problems and deficiencies in this region, and some of these issues have been increasing in the last decades in a very alarming manner. Although there is good access to public and private services, which makes the metropolitan zone very competitive against the vast majority of Latin American countries, the inadequate and unfulfilled territorial and urban planning has generated major problems in terms of public infrastructure, transportation, pollution, social insecurity, and depopulation trends in the center of San Jose, the capital of Costa Rica.
As pointed out by the Ministry of Housing and Human Settlements (MIVAH), the urban development structure of the GAM has been historically radial, spontaneous, and disperse, often not complying with the stablished urbanistic norms; a so-called “sprawl” urban expansion [15]. This not only affects the infrastructure and transportation routes of the region, but entails a set of negative social, economic and urban implications. In addition to this, the lack of housing and green areas in San Jose directly affects the attractiveness and competitiveness of the GAM as a center of investment and tourism.
Currently, one of the most impacting issues affecting the GAM is vehicular mobility, both in terms of the amount of vehicles on the streets, as well as the time it takes to commute from one place to another. According to the State of the Nation Report from 2014, Costa Rica holds one of the most dense road infrastructures in Latin America; statistics that have remained for several decades now, with a tendency to get worse each year [16]. The country’s infrastructure is practically collapsed in terms of functional capacity and without any physical capacity to expand. Furthermore, as the urban development is focalized in the central area of the country, the volume of vehicles in the GAM is extremely high, creating an even bigger and more difficult issue.
Additionally, Costa Rica has a very deficient public transportation system; there are short and long-distance buses, and just a limited number of inter-provincial, non-fast trains. According to the Urban Observatory of the GAM (OUGAM), buses are the most common mean of public transportation used by the citizens of the country – around 57.2% of them use this service-, although this fact is rapidly decreasing lately as more people are opting for using their private vehicles -approximately 45% of the families in Costa Rica own one- [17]. Several factors are influencing this, including a system that gives priority to the use of private vehicles over the massive transportation of people, as well as the time saved by using a private option compared to using public transportation.
Many of the current issues in the public transportation system lie in a practically inexistent schedule for most routes and a lack of infrastructure, planning, and information systems that allow a better administration of the services provided. Coupled with this, there is
a high demand for transportation in such a limited physical space, especially because most roads do not have specific lanes for buses, which sums up to the amount of vehicles circulating in the GAM. But why are all these issues affecting so much the mobility of people in Costa Rica? What are the internal and external factors that influence the whole context of this problem? In order to answer these questions there is a need to comprehend the problem beyond the traditional definition of mobility in a modern city; there is a need to incorporate governmental, social, economic, environmental and technological elements. This way the problem statement can be understood, a solution (or multiple combined solutions) can be proposed and results can be perceived by the population in the mid-long term. Mobility issues affecting the GAM are so vast and have been so intrinsic into the costarrican life for decades, that there is almost no short-term solution for this, at least not one that can fulfill and improve mobility in a small amount of time.
Government
Costa Rica is administered by three main organs (Executive, Legislative, and Judicial), all of them being completely separate and independent from each other in their functions and authority. This form of governance was based on the political principle of separation of power, trying to address the high level of corruption and management issues that have been affecting the government of Costa Rica since many decades ago. This type of division of responsibilities is, in theory, a good approach to address many of the problems that can occur when a small group of people leads a country, but it requires really high coupling, communication and efficiency between the parts to really operate in a way that the country can benefit from the model. Unfortunately, this is not the case in Costa Rica, this separation of duties have constantly led to an absurd amount of bureaucracy in public institutions, constant disagreements not only between the state organs, but also between the parties managing these institutions, and an overall inefficiency of the Government when trying to implement changes or create new solutions for the problems affecting the country.
Specifically in terms of urban mobility and territorial planning, Costa Rica has “a very slow continuity” as stated by Alfaro [18]. “There is a great problem for the GAM and for Costa Rica in general, as the region does not own any competencies conferred by law, there is no technical organ, nor a regional institution that administer the functions of the territorial and urban mobility planning in a regional or provincial level”. This means that any issue affecting the mobility and territorial planning of the GAM must be directed to the upper levels of the Government, or to each cantonal municipality, and there is no dedicated entity foreseeing to plan, execute and control any of the operations related to the urban mobility and growth in the GAM. This clearly leads not only to an uncontrolled and chaotic administration in the area, but it also results in unbalanced and independent efforts among the different municipalities of the GAM; San Jose for example, has been through a great effort to improve the city during the last decade with fairly good success, but the rest of cantons in the region have a significant development backwardness in terms of urban development and mobility planning, affecting the GAM as a whole.
Furthermore, even when San Jose has had a noticeable urban development during the last 11 years, many of the core issues affecting the city still persist; Fornaguera described San Jose in 2013 as a city with “increasing homelessness, pollution, chaotic transportation, but with green spaces, boulevards, festivals and expensive condominiums” [19].
The Research Program for Sustainable Urban Development (ProDUS) of the University of Costa Rica also studied the fact that the small amount of housing available in San Jose has created a “dormitory town effect” in the GAM, as many people work in San Jose but most of them live in the peripherals of the city, creating a need to commute to and from San Jose in a daily basis [20]. This with no doubt, instantly translates into more pollution, more time lost in traffic jams, greater fuel consumption, losses by the many car accidents that occur daily in the GAM, and an overall increase in the cost of goods and services.
Despite of the existing and recently proposed transportation reforms, the creation of decentralized institutions such as the National Roads Council and Public Transportation Council, and many other infrastructure projects, most of these efforts remain in paper only, they have never started, or show progress in an incredibly slow rate. As stated in the State of the Nation, most of these efforts are affected by poor planning, inefficiency in the processes, excessive bureaucracy, a debilitated technical workforce and scarce financing [16]. All these aspects together result in projects that take years to start and even more time to conclude, if they actually manage to get to a conclusion.
As stated by the Inter-American Development Bank (BID) in 2014, Costa Rica needs a strong interaction and cooperation between the all parties from the government and more investment from local and international entities [21]. Moreover, these aspects require a better and more efficient communication, and specially a governmental, technical and proactive mechanism that assures that the projects are actually planned for the long term, executed and constantly maintained. It is also stated by the BID that any effort focused on massive transportation systems and territorial ordering are of utterly importance and must be prioritized in order to influence a better urban and transportation administration.
Socioeconomic and Environmental Factors
As defined by Rozas & Correa, suburbanization, fomented by an improvement in the household income, consists in the displacement of families towards zones where they can get rid of their dependency of public transportation, primarily due to the acquisition of a private vehicle [22]. Generally speaking, this phenomena starts with people living and working in or near the city center, using public transportation to commute, and only using their private vehicles (if they own one) during the weekends or at non-peak hours. Subsequently, they as their economic capabilities raise, they leave the city center and move to external suburbs and their need to use their private vehicle increases, resulting in a progressive independence from public transportation, which at the same time causes an increase in the vehicle fleet transiting in the streets. This occurrence has affected many Latin American cities during the last decades, carrying with it a lowering in demographic density, but at the same time forcing to improve the roads system both in its condition and its capacity. As a result, according to the Sectorial Planning Office, just between 2007 and 2011 the amount of kilometers traveled by vehicles in the GAM alone raised in more than 3 million per year [17]. This is caused by a great increase in the amount of vehicles, and the distances that the users need to cover to commute on a daily basis.
It might seem that expanding the capacity of roads and highways, and investing in maintaining these systems could be a good option to improve mobility in a city, but this is not certainly correct, at least not for the GAM. This model only incentivizes the use of private vehicles, and requires a join operation between all levels of roads administration and planning, which, again, is not the case in the GAM. Many of the latest projects performed in the region (i.e. new highways, overpasses, extra lanes in roads) are isolated efforts that try to solve the problem in just one small fragment of the road system. This, ironically, results in an even bigger mobility issue, as many more vehicles transit by these new sectors, but end up again in the segments that have not being improved, causing enormous traffic jams and many accidents. Doubtlessly, these issues are also promoted by a generalized costarrican culture of not caring much about traffic laws and a great disrespect for other drivers, which is also exacerbated by the distress and despair generated by the many issues that they have to face every time they have to transit this problematic road system.
These socioeconomic and governmental characteristics create a negative vicious circle that affects the GAM as a whole. According to Arango, when people travel using their private vehicles instead of a public transportation system that is not attractive, this automatically translates into more vehicles on the roads, therefore more traffic jams, greater traveling times and more pollution and environmental issues affecting the region [23]. Moreover, as these problems grow bigger each year, the public transportation usage is more discouraged and provoke more urbanization on the outskirts of the city, generating, again, even longer traveling times and more issues affecting the mobility of the entire GAM.
In terms of environmental impact, the pollution produced by the transport subsector represents more than 60% of the total contaminating emissions of Costa Rica as pointed out by Kowollik in 2014 [24]. Because of the poor railway infrastructure existing in the country, most of the transportation of goods and passengers is done by land using the inefficient and problematic road network. Besides the amount of private vehicles circulating in the GAM, the bus companies (which are private and independent from the government) are required to renew their operative licenses every seven years, but these licenses do not include any environmental exigencies, resulting in poorly maintained units, creating a huge amount of pollution, CO2 emissions, and creating an even more chaotic situation.
These facts not only affect the economy and the environment of the region, but also negatively influence the Costa Rica’s goal of becoming a carbon neutral country by 2021. As stated by the Ministry of Environment and Energy in their 3rd Nacional
Communication for the UN Convention about Climate Change in 2014, the last inventory of CO2 emissions in Costa Rica exposes that more than 8.78 millions of tons of CO2 were released in 2010 [25]. From these figures, about 4.53 million of tons came from ground transportation alone, which translates into having more than half of the total emissions of CO2 in Costa Rica coming from vehicles. The subdirector of the National Institute of Meteorology, Roberto Villalobos, exposes in the document that “people from Costa Rica not only prefer to use their own private vehicles instead of public transportation, but they are not accustomed to share their vehicle with other people, resulting in higher pollution and emissions derived from private vehicle usage”.
Technology
Besides of the previously stated issues, one the most important, but often excluded factor in the mobility of the GAM, is the use of technological devices, systems, and platforms that could help to improve the efficiency, administration and maintenance of roads and vehicles circulating through this area. Despite of the few efforts put into creating new solutions to increase the mobility in Costa Rica, most of the public infrastructure, both physical and technological, dates from the 70’s decade as mentioned by Kowollik [24]. This is illustrated by deteriorated buses for most national routes, non-digital payment methods, inexistent schedules and almost no systems or devices to administer roads, traffic laws, parking places and many other issues that are closely related to technological features. The main characteristic of this, is that those aspects could be technologized, but only if the government and people actually focus on this type of solutions and leverage them to solve at least some of the issues affecting the GAM.
Recently, there have been a few efforts to include digital payment platforms for the public transportation system, with a very slow, not very popular and almost no materialized deployment of the platforms; even when this a trivial part of the operations of modern transportation systems. Most of these projects have remained only on negotiations with the government (many state that this happens mostly because of conflict of interests between the parts) and only a small pilot testing was done with only one of the more than 40 bus companies operating in San Jose.
The platform called PagaFacil is currently being tested using buses of La Periférica (a bus company that operates throughout the peripherals of the GAM, hence its name). This platform still remains in its early deployment phases, mainly due to a lack of investment, many bureaucratic processes and absence of promotion from the government. PagaFacil is an example that it is required not only to have the technical capabilities to develop a system, but there is a great need of encourage usage among its users, especially as the costarrican population is a very skeptic of using these type of new technologies. Ricardo Herrera, owner of PagaFacil, stated that the government should also assist these type of efforts as most of the time it create more impediments instead of supporting them. Moreover, technology should not only be included into the problem equation because it provides benefit to the mobility in Costa Rica, but it should be part of any proposed solution as we live in an era in which information, and its administration and usage, is an essential part of our lives; a tool that is sadly being squandered in Costa Rica. Nevertheless, platforms like Uber have created an incredible impact on how technology is perceived by the costarrican population, providing a technological approach to have a cheaper and more efficient service, but also fighting the corrupt and poor public taxi service operating in the country. Another important aspect of this application for the costarrican culture, is that it is removing the fear of using online applications that manage private information, such as credit card data; this is something that most of the population was historically skeptic to do, especially older generations.
This is what new technological solutions should do in populations like Costa Rica, not only by providing a great service, but creating empathy and security in users so they can be certain that their data is safe. Part of the success of these platforms are tightly coupled with the characteristics of a sharing economy, proving that this type of services are part of the new technological revolution that is happening in a global scale.
5.2. Mobility for TEC Students
Part of the context statement of this research was to find how students from the Costa Rica Institute of Technology commute to the university. This was important, first, to correctly design and develop the application prototype that was tested with these students, but also to provide an understanding of their mobility needs, and what was their perception about deploying the pilot application in this institution.
This part of the study consisted in a personal interview with 100 randomly chosen TEC students. The conversation consisted in a set of questions that foreseen to understand how these students commute. The first group of questions were directly related to discover what means they used, in what frequency, and the average time used to travel to the university. Then, they were asked about which places they usually travel from and to, and which routes they took. Lastly, the interview consulted the interviewees if they currently share their vehicle while commuting -or if travel with someone that does it-, and if they were willing to use a carpooling application with other students from the university. The questions and results of this phase of the investigation are available for online consultation [32] and are going to be summarized next.
First, it was exposed that 65% of the interviewees used public transportation to commute to the university, 16% used their own private vehicles, and 9% stated to use both. Moreover, 10% of them did not use any of those two options as they lived near to the TEC campus, which allowed them to walk or use a bike to travel to the university. Then, from the 90 interviewees that commuted using a vehicle, 73.3% stated that they travel on a daily basis, 17.8% do it every week –mainly because they rent an apartment near the university-, and 8.9% in longer periods of time. Then, their responses showed that 42.2% spend less than 30 minutes to get to the university, 21.1% of them travel for approximately one hour, and 36.7% of the students commute for more than one hour. From this first group of questions, it was observed that the amount of students using public transportation closely matches the numbers provided for the country’s statistics provided by the OUGAM [17]. Then, the high frequency of daily commuting students resulted in great feedback for the development of the prototype, as it exemplified that there were many people traveling often to the university. Lastly, the different results in the duration of these journeys was also positive for the research, as the carpooling prototype could be designed a way that it satisfies the different types and durations of the rides.
One of the first designs of the prototype was based on fixed stops or meeting points, where drivers and riders could meet and then travel together; the idea was to define these points based on the most frequent and popular places where students commute from. Nevertheless, the second group of answers exposed so many different places and routes, that the idea had to be discarded. Therefore the new design was based on places that the users could define and pick freely, both for the place of arrival and destination. This model at the end, resulted in a more open and scalable design, as future versions of the application will not be constrained by the places that only TEC students use.
The last group of questions revealed that from the total of students owning a private vehicle, only 38.5% of them did not share it while commuting, and from the remaining 61.5%, 23.1% do it every time they travel. These results were very positive to the overall research as they showed than more than half of the interviewees already shared their vehicle or travel with someone that does it, even when
there no existing system or application to help them to do it. Moreover, an impressive 98.8% of the students stated that they would be willing to use a carpooling application for the university and travel with other students. In other comments, the most mentioned aspect was the need to generate trust for both users and drivers; many interviewees stated that they would like to know who they are going to travel with in advance, and that there should be a mechanism to assure that only students of the university could use the application.
In summary, the results of this phase of the investigation were of utterly importance not only to define the mobility characteristics of TEC students, but also to have a design based on the needs and concerns exposed by the users themselves. This user-centered design approach set the basis from which the rest of this study was performed, and from which the design decisions of the system were taken. All these design features are thoroughly detailed and exemplified in the System Description of this research.
6. SYSTEM DESCRIPTION
In this research an Android application called Iter (which means journey in Latin) was developed in order to test and evaluate the requirements obtained from the theoretical exploration of this investigation. The main purpose of the system is to encourage carpooling usage among the students of the Costa Rica Institute of Technology, allowing users to search, post and reserve rides, or Iters as they will be referred to from now on this document. Moreover, because of the good results and feedback obtained from the first part of the research, the system was designed in a way that it could be expanded and used in other universities and institutions in the near term.
The following section first describes the architecture components that constitute the system, explaining the reasons why these components were chosen, and how their joint operation helped to create a high-end solution in a fast, economic and efficient manner. Then, in System Design, the frontend and backend of the system are described in detail. The frontend subsection includes an explanation of the patterns and principles used for the graphical user interface (GUI), and how the different screens and functionalities of the system were designed. This is followed by a description of the backend layer operations: how the system along with the external libraries and APIs allow the application to store, retrieve and process the information used throughout the platform.
6.1. System Architecture
The system architecture is mainly based on three large components: The user’s mobile device, the platform’s server, and the external systems. Following, these components are broken down into smaller subcomponents in order to describe how they operate and work together.
The architecture diagram (Figure 1) illustrates the high-level communication between the client device and the backend layer, which is composed by the platform’s data management server and the external systems and libraries. Every component of the system is given a specific task or role to undertake, and this division provides the ability to improve each component separately. This distribution of components is based on a modular system approach; allowing better scalability, maintenance and improvement of any of the components of the system without critically affecting the other parts.
Figure 1. High-Level Architecture Diagram
When an interaction is initiated by the user, the client device communicates with a request to the respective backend component resulting into a response, which is again managed and presented to the user in the GUI. This model constitutes a simple client-server communication approach, nevertheless, the combination of both internal and external backend components results into an easy, efficient and transparent process, benefiting the overall interaction between users and the system.
Architecture Components
The following subsections summarize and provide a description of each one of the components included in the system architecture, and the interactions between users, devices and the backend layer.
Client Device
The client device represents the piece of hardware in which the application will be installed and used by users. This research found out that the most used operative system (OS) in Costa Rica was Android [26], therefore, a mobile application for this OS was built. However, the fully operational system is not fixed to this operative system only; iOS, Windows and the Web, are certainly alternative options to deploy the application, and can provide a more far-reaching set of users. The main hardware requirements for the client device are internet connectivity such as Wi-Fi or mobile data, local storage available in order to save the user’s session, and optionally, location capabilities, such as GPS, to provide a better experience. The device is responsible of sending user requests and processing the responses coming from the backend layer. These requests are generated from the interaction between the user and the different functionalities of the system that will be described in System Design. As the architecture is modular, is part of the responsibilities of the device to have the logic to communicate to the corresponding backend layer component in a way that is transparent for the users; meaning that, if any backend component is modified, the users should not notice any difference in the interaction, allowing a more consistent and efficient way of interaction with the system.
Platform Server
One of the key elements of any client-server architecture is the backend, this component is the server (or servers) where the application data is going to be stored and retrieved from. For the system described in this research, this component was essential to store all the information about users and Iters (location, date, description, etc.), and other necessary data as it is going to be fully detailed in System Design. Additionally, and according to the requirements extracted from the first part of this research, it was necessary that the data management server used in the platform provides good response times, strong security features and independence from the type of client device used by the users.
There were many approaches that could fulfill this model; they range from locally-stored information -using the device storage, but resulting in decentralized and very volatile information-, to privately hosted and managed servers -with high costs to acquire and maintain them, but providing more efficiency and total control of the data stored in them-. Both of these approaches had pros and cons, which forced this research to find an alternative solution; something with the benefit and efficiency of privately hosted servers, with low costs of acquisition and maintenance, always taking into serious consideration the future scalability of the platform. After evaluating the few existing options with these characteristics, Backendless resulted into the most appealing and most technologically favorable solution; it is a free, cloud-based Mobile Backend Service (mBaaS), with great data storage capabilities, and an easy to use web-based management console (Figure 2) containing analytics, and many other features that will be detailed later in this document [27].
Figure 2. Backendless online management console dashboard
Backendless fulfills all the hardware and functional requirements of the proposed architecture model, both in terms of costs and efficiency. But even with these features, it was of utterly importance, due to the modular characteristics of the architecture, that this data service could be replaceable with another type of service or server; in order to achieve this, good migration capabilities are needed, again, Backendless accomplishes this purpose with its data migration functionality, making this service the best option for the purpose and goals of this research.
Mobile backend services are relatively new in the industry, but are part of the most recent agile practices that are being used throughout many applications and services all over the world. This approach fits perfectly to the purpose of this research and represents a very accessible and efficient way to develop a system in a short period of time; features that allowed to make the prototype of this research in about two months, including design, development and deployment.
External Systems
The third architecture component is tightly coupled with the platform’s backend, they are the external systems and libraries that complement the backend layer of the system. This component is constituted by services and functional libraries that help to satisfy the requirements of the application. Most of these services and libraries could be manually implemented instead of depending on external providers, but this would represent much effort in time and development, especially when it is something that is already done and available to use straightaway. For this research three main requirements were identified as external services or libraries.
First, it was needed to have a way to authenticate users in the application, for this, Facebook API was chosen to handle these accounts, as many modern services and applications opt to use this service for its easy and straightforward setup requirements [28]. Backendless also has very good integration capabilities with this API, which makes the coupling of both services very easy to code, configure, and maintain. Moreover, many users stated during the qualitative evaluation of this research, that Facebook is an easy, trustworthy and fast way to log into the application.
Secondly, as the prototype was planned to be used only with TEC students, one important requirement was to allow users to register their University ID. In order to achieve this, the research first evaluated the possibility of using the existing TEC databases and systems, but there were many permissions needed and many technological limitations for this process, therefore other options had to be considered. Then, the research found out that using BlinkID the application could read and store the barcode used in the students’ ID card. This library consists of a set of image processing and recognition functionalities that identify many types of IDs and barcodes, allowing to use these features with just a minimum coding [29]. Then, the library was configured so the prototype application could only read ID cards from TEC students, nevertheless, it was also coded in a way to allow future modifications to the library in order to read other types of IDs. This was especially important as the application is planned to be scaled up and used in other institutions and companies in the near term. Lastly, one important element used in the system was the geographical information used for travel directions, and pickup and destination points. For this purpose Google Maps API came as the first option to accomplish this for its free and easy to use functionalities [30]. Furthermore, many users are very accustomed to this service, creating better usage consistency and ease for them, aspects that are very important to take into consideration when it comes to geographical information, as it can be difficult or confusing for first-time users. As it is going to be explained in System Design, the functionality to search and pick a place in the map is not only the same as used in Google Maps, but it is also used by many other geo-enabled services such as Uber. Again, the main purpose of using this API was to include a standardized and known process, and especially because of the time saved by using a third-party provider for geographical information, instead of trying to develop those features in a manual way.
Architectural Operation
The previously described system architecture is not only intended to provide clarification to the reader of the components used in the application prototype of this research. It also pretends to help and motivate other researchers, developers and designers to use it as a model. The architecture is designed in a way that other systems and applications can leverage the capabilities of this modular approach, and serves as a proof that the agile methodologies, components and services of the architecture can help to develop a high-end solution in a relative short period of time. Even when some of the components previously explained are mainly related to the developed prototype, the idea behind this architecture is to demonstrate that the joint operation of both internal and external components and services could really benefit the development of an application or system. This cooperative integration of the parts not only reduces the time needed to create a solution, but it also helps to provide the users with functionalities and processes that they are already used to work with, improving the overall ease and efficiency of the system.
6.2. System Design
This section describes in detail the design decisions taken to create the application prototype of this research. It is divided in two main sections: Frontend Design focuses on explaining what design principles were used, and how the interactions and functionalities provided on each screen of the prototype work. On other hand, Backend Design describes the model used to store, retrieve and process the information used in the prototype, it also includes how the internal and external components operate together to accomplish a seamless and efficient operation.
Frontend Design
Frontend refers to the user-facing portion of the system that provides users not only with the GUI, but also with all the interaction and functional capabilities of the system. First, this section defines the Google Material design principles used throughout the system and why they were helpful while designing the application. Then, the different screens or sections of the application are explained, including every functionality available in them, and how the interactions with the user were designed in order to promote an easier and more efficient usage.
Google Material Design Principles
Google Material design principles [REF] provided a great source of information for the decisions made in the prototype of this research. As the application was required to be built on an Android operating system, these guidelines allowed to build a GUI following market proven design principles for this specific operating system. The principles are presented in a complete and easy-to-follow guide; starting from the core concepts of material properties, later introducing style, layout and component characteristics, and finalizing with design patterns along with usability principles.
Google’s principles generated good consistency throughout the application and a better designed interface, resulting in a more intuitive and easy to use system. As explained in these guidelines, if an application is built using random guidelines, styles, and layouts, it will generate the same feeling of randomness for the users while they use and experience the application. For example, if an application has an uncommon layout, the user will first have to learn and understand the new distribution of the elements on the screen. By using layout patterns, this will save effort for the user and they will be focused on the content rather than on the elements of the GUI.
These patterns not only help designers to build better experiences for their users by providing and explaining how and why to use them, but also by explicitly defining common and frequently used design decisions that generate issues. The ”do’s and don’ts” recommended by Google vary from good interaction techniques, style decisions, such as using bad font faces and sizes, to specific layout details, such as where to position elements on the screen and how to use material characteristics, such as thickness and depth of elements. In general they provide a solid foundation to create visually appealing and intuitive interfaces in an easy manner. Specifically, there are three main features that helped to create a more visually appealing and efficient prototype: Elevation and shadows provide a good way to create the feeling of working in a 3-dimensional space. Interaction responses provide the user with visual feedback of any interaction occurring in the interface, such as an expanding circle while clicking a button. Meaningful motion,
serves to focus the attention and maintain continuity between transitions in a subtle yet clear manner, this was used between screen transitions, for example while opening an Iter, expanding the user image while it loads the Iter details, helping not only to have a good visual feedback, but also giving time to the elements to load instead of making the user wait with a blank screen.
Welcome Screen, Login and Registration
The first page that the users will see right after they open the application is a welcoming screen (Figure 3), followed by the login and ID registration page. This screen (or activity as they are called in Android) contains a View Pager, this is a tabbed component that allows to scroll horizontally through its views. The objective of this component is to guide the user through a step-by-step wizard, in order to login and register the user’s ID.
Figure 3. Iter Welcome Screen
These three views are designed with a big icon in the middle of the screen representing the action that should be taken on each step, and a description below the icon explains the purpose of each section. The welcome screen uses orange as this is the primary color of the application. The login screen uses blue according to Facebook’s primary tone. ID registration uses green as this is the secondary color used in the TEC pallet (primary is blue).
To login, users are requested to use their Facebook’s username and password, the component used for login is automatically generated by Facebook’s API, and all the logic behind it to check if the credentials are valid belongs entirely to the API. Iter does not use or store any of this information, first because it is technologically impossible to retrieve these values from the component generated by the API, and most important because it represents a great violation to the user’s privacy. If the credentials are valid, the application then checks if the user account already exists in the backend and the account information is retrieved. If the user does not exist, a new user account is created and stored in the backend. In both cases the account information is stored locally in the device to avoid to login again when the application is closed.
After the login is complete, the system checks if the user has already an ID associated with the account, in a positive case the user is then redirected to the main page. For first time users they will need to register their university ID card, as described in a previous section of this document, BlinkID provides all the recognition capabilities to read and store the barcode that is on the ID. Figure 5 illustrates how the ID registration process looks like; the library uses the device camera to detect, capture and process the barcode in an easy and fast way. After the ID is registered and associated to the user’s account the main activity is displayed.
Figure 5. BlinkID Registration
Main Screen & Navigation Drawer
As soon as the user completes the login and registration process, it is redirected to the main activity and the navigation drawer is automatically opened (Figure 6). The objective of opening this component is to teach the user about the existence of this drawer and it only happens the first time that the user opens the app; this feature was implemented as part of the Google Material guidelines. The header of the navigation drawer displays a big picture of the user that is retrieved automatically from the Facebook profile. The name and ID number is also shown in this section to display visual feedback that the login and ID registration process were executed correctly. The next section of the drawer is related to the Iters; the first option allows to create a new Iter; the second group of elements allow to filter by available Iters or by “My Iters” which are those that the user created or previously reserved. Next, the score of the user is shown, which represents the feedback given by other users, similar as used in many similar services as Blablacar. The last option in the drawer is a logout button which not only logs the user out of the app, but also removes any preference and data stored locally in the device.
The main screen (Figure 6) is composed by two elements: The main toolbar, which contains three buttons, one for showing the navigation drawer, a search button which will be described shortly and a button to create a new Iter. The second component is a Recycler View nested within a Swipe Refresh View. The recycler is the component in charge of creating each one of the Card Fragments that represent every Iter stored in the backend. These cards are part of the Material Design guidelines and provide great memory management features. It also creates an efficient data retrieval process, as it uses data paging while obtaining the Iters; it obtains more available Iters only as the user scrolls down, saving not only memory and processing power, but creating smaller and more efficient requests to the backend. The swipe view allows to refresh the existing Iters by pulling down the top part of the screen, as used in many popular applications nowadays.
Figure 6. Iter Main Screen & Navigation Drawer The list of available Iters displays only future Iters and they are ordered by date; putting the ones closer to their departure date on top. Each card shows a summary of the Iter using three main pieces of information: the person that created the Iter, his picture and score, the meeting point and arrival place, and the hour and date when the Iter departs. Only this information is shown, as putting more data will overload the card; this layout also encourages users to open the Iter details if they are interested in it. Moreover, using a small amount of information in the card helps the data to be correctly displayed in smaller screens.
Using the search functionality the user can retrieve any Iter that is between a range of 1 and 50 kilometers away. Google provides with specific guidelines about location retrieval in order to preserve battery and memory of the device, because of this, the app was designed using a mix of GPS, Wi-Fi and cellular data to determine the user’s location.
Further functionalities of the application should allow to search Iters by departure or arrival place, using specific date ranges or even the owner of the Iter. These functionalities were not developed in the prototype of this research as they were not included in the system requirements, and did not represent any primary functionality that needed to be included in the pilot testing.
Iter Creation
The process of creating an Iter consists of two simple steps: First, the user is asked to provide a meeting point –here is where the user is going to pick up any person that reserves the Iter-, and a destination, followed by the time and date of departure.
Google Maps Place Picker is the component that allows the user to choose the meeting or destination point. There are four ways to choose the desired place: Dragging and moving the pin on top of the map, tapping on any of the red dots, using the list of nearby places or searching for a place writing its name. As it was stated before, this mechanism is very similar to the process used in Uber, Google Maps, and other geo-enabled applications. One of the main advantages of using this feature is that Google’s API has great functionalities to obtain the name of places, streets and cities, allowing an easy and faster identification for the users.
Once the information about location, time and date is complete, the user is then asked to provide the details of the Iter. For the prototype of this research prototype only two piece of information were necessary: the amount of free spots on the vehicle, and a small description that include any useful information for the people reserving the Iter such as car model and color, schedule flexibility, etc. Once all the information is complete the user can click on the Create button and the Iter will be saved in the database. If any information is incomplete the application will alert the user and ask to complete it.
For the purposes of this research using the description field was enough to fulfil the need to include the details of the Iter, but for future versions, this section will include more specific fields such as cost of the ride, the license place and photos of the vehicle, among other details that could be relevant depending of future user requirements and functionalities.
Iter Information and Reservation
When a user selects an Iter from the main screen, all the details of the Iter will be displayed in a separate screen (Figure 7). This page has a scrollable and vertically oriented layout, it starts with a big photo of the owner of the Iter in order to provide visual security feedback to other users. The name of the owner, the ID and the score is also included in this section to offer even more trust. Following, the Reserve button is shown and it could be used to easily reserve one spot of the Iter (if there are any available). Then the details of the Iter are presented to the user, including departure and arrival, time and date, available spots and the description provided by the owner.
The last part of the screen displays the list of users that have already reserved the Iter, this is helpful to provide even more confidence to the user as everyone can see the name, picture and score of the people that they will travel with.
Extra Functionalities
Future development will allow to click on the departure or arrival place of an Iter to open Google Maps, showing recommended routes, estimated travel time, among others, in order to provide an even more complete and useful service. Moreover, in future versions of the system, every time a user reserves, or if the owner makes any modification to the Iter, every person traveling with it should receive a notification in order to be informed of the latest status of the Iter.
In other aspects, even when the application shows rating scores for users, this functionality was used only for testing purposes of the prototype. The rating system of the application should let users score both the driver and the people that reserved the Iter after it finishes. Uber uses this system in an easy-to-use way, providing with a 5-start scale, which can be easily averaged and assigned to users. This system along with functionalities to connect with Facebook friends provide a more complete social experience which directly translates into more user confidence.
Moreover, the use badges when milestones are achieved (i.e. after completing 10 successful Iters, or by traveling with more than three people) could also provide a great form of gamification to enable even better user engagement, trust and satisfaction by using the app.
Figure 7. Iter Details
Backend Design
In this section the model used to store, retrieve and process the information used for Iters and their users is detailed. Furthermore, it explains how the different components work together to accomplish the system requirements and the functionalities described in Frontend Design.
Every group of information specified in this section, is stored in the Backendless data repository using a regular client-server communication protocol as explained in System Architecture. Besides the data that is going to be defined next, Backendless automatically stores for every object a unique identifier, the owner (or author) of that object, the date it was created and the last time it was updated. Additionally, for users it stores the account status, the last date of login and the type of social account (which is always Facebook in the prototype of this research, but could be a Twitter, Google or Microsoft account). This is a great functionality as it is not needed to code or configure anything extra in order to leverage the use of this information in the application. Furthermore, this data is automatically refreshed as the backend objects are stored, retrieved or updated, providing even lesser need to code or update it manually.
