This study was done at the Eastern Africa National Networks of AIDS and Health Service Organization (EANNASO) in Arusha-Tanzania. In addition to EANNASO, further research was conducted in other CSOs located in Dar-Es- Salaam and Dodoma regions in Tanzania. The choice of these regions for the study is because they are major cities and have a substantial number of CSOs according to data from the Foundation for Civil society [8] as depicted in Figure 1.

During the stage of identifying the problem, a literature review was first conducted to understand the current situation on how ICT is leveraged by players in the non-profit sector i.e. CSOs, Community Groups, NGOs etc. in streamlining

their operations and strategic objectives.

In the process, physical interviews, questionnaires and observations techniques with 140 players in the non-profit sector in Arusha, Dodoma and Dar- Es-Salaam regions were conducted. The participants comprised 72 males and 68 females. The participants had an average age of 35 whereby 45% of them came from CSOs, 35% from Community groups and 20% were individuals who are in the marginalized category such as the elderly, the disabled and unemployed youths.

In the interview sessions, questions such as 1) Have you ever wanted information about activities of a CSO/Community Group but did not know where to look? 2) Have you ever needed contacts of CSO/Community Group and failed to get one? 3) Have you ever obtained information about a CSO/Community Group but later found out that the information is outdated or inaccurate? 4) Have you ever tried to find physical location of CSO/Community but failed? 5) If you are a leader/owner of a CSO, have you ever wanted to publish/publicize your CSO and its activities but failed to find a suitable platform to do so? 6) In your opinion, how can we enhance accessibility to information about CSO/Community Groups? 7) Do you think a mobile application with information and location about CSO/ Community Groups will help to solve some of the challenges you are facing? The feedback of the interviews was recorded and then, analyzed through the use of inductive coding approach. The analyzed feedback were tabulated to show the frequency of responses from respondents in order to determine the most persistent problems facing the current landscape of CSO as far as information is concerned. This in turn would help the study in devising the solution to address the major pain points.

Rapid Application Design (RAD) methodology has been used in the development of the solution to the problem identified in this study. RAD was used, because it offers iterative development with high quality result of information systems [9]. In the design and development process, the study conducted a total of five iterations in which features such as UI/UX design, application speed, scalability, security and robustness were evaluated and revisited as per user feedback on the prototypes. The prototypes were developed in Android/Java/Kotlin (frontend) and NodeJS/MongoDB (backend). NodeJs was chosen for the backend because of its high speed in serving API requests since it uses event-driven architecture and has non-blocking IO (Input/Output) [10]. MongoDB NoSQL database was chosen because it has more scalability and efficiency in storage and access of data than traditional relational databases like MySQL [11]

In the system evaluation phase, a field study was conducted using the final prototype. The field study consisted of 60 diverse individuals who are players in the non-profit sector in Arusha, Dar-Es-Salaam and Dodoma regions. The participants were 36 males and 24 females whereby 55% were from CSO, 25% from Community Groups and 20% from marginalized groups (elderly, disabled, unemployed youths etc.). The questions asked in the evaluation study were on the assessment of the developed system if it has enhanced the access to information about CSO/Community Groups to interested stakeholders or not. Other questions focused on the ease of use, security, privacy and interoperability of the developed system and then, the results were processed using Pandas library of the python programming language.

The functional and non-functional requirements collected from the users for the proposed system are summarized in Table 1. The requirement data was collected by using questionnaires, observations and conducting interviews with different stakeholders in the civil society arena.

The system has been designed following the client-server model. Based on the analysis of the requirements, the server (backend) has been designed following the Model View Controller (MVC) architecture because of the inherent benefits of this model such as enabling scalability, maintainability and robustness of software. Additionally, MVC makes reuse of code possible and enables different developers to develop the system in parallel. With MVC pattern, parts of the system can be modified without affecting the system as a whole [12].

The frontend has been developed in Android using Java/Kotlin following the Model View ViewModel (MVVM) pattern which is the recommended architecture for building robust, production-quality applications. MVVM architecture

implements the principle of Separation of Concerns between the different modules of the mobile application. Separation of concerns achieves modularization of the application making it easy to scale, read and unit-test the application [13]. Figure 2 shows the architecture of the whole system i.e., frontend (MVVM) and backend (MVC). In the frontend, the Activity/Fragment is responsible for presenting the UI to the user. The ViewModel is responsible for providing data to the UI. The Repository is a mediation layer that selects whether to pull data from the local database or making an API call to the remote data source (webservice). The SQLITE database is a local store of data on the mobile device. The transfer of data between the mobile frontend and the backend is accomplished using JSON (JavaScript Object Notation) format. In the backend the Controller is responsible for receiving and sending responses to API calls. The Model is a schema describing the structure of data and the View is a presentation layer that displays data to the user.

The following are the features of the proposed system

1) An android mobile application which will be responsible for providing a UI/UX for the user to interact with the system. The mobile application renders a view that displays all the CSOs in the system. It provides a form that users can use to upload information about CSOs. In addition, it senses the user’s location (after requesting for permission from the user) and displays map view that shows all CSOs. The mobile application also provides search functionality.

2) A backend which is written in NodeJs using express framework. It is hosted in a cloud application hosting service called Digital Ocean. The backend implements a RESTful API that provides routes that serve HTTP requests from the mobile frontend in the JavaScript Object Notation (JSON) format.

3) A database for storing CSO information. The database used by the system is a MongoDB NOSQL document based database. This system uses MongoDB Atlas which is a database as a service (DbaaS) hosted in the cloud on Amazon Web

Service (AWS). Images and other media of the system have been hosted on Cloudinary platform which is a cloud service for hosting and transforming images and videos.

4) An administration panel that is web based which is developed using ReactJs frontend library. The ReactJs library has the benefit of saving time of development and minimizing the effort to maintain code [14]. The panel enables an administrator who has full rights of create, read, update and delete (CRUD) as well as to manage users and content of the system.

Figure 3 shows the conceptual framework of the developed system.

Use case diagrams are used to show how one or more actors interact with the system. Use case diagrams depict what the system can do, i.e. how user inputs are handled in relation to the system’s functional requirements [15]. In this application three actors were identified i.e. user, admin and Oauth Provider.

Users of the system interact with the mobile frontend. They can pull all data about CSOs, they can search data according to key queries such as country, region, CSO focus area etc. In addition, CSO/Community stakeholders can register to the system and upload information about their particular CSO by filling a form. The data is sent to the backend and stored securely in a cloud based MongoDB database via a NodeJs backend. In addition users can also geo-locate CSOs in an interactive Google map. The map also integrates Google’s direction API that enables users to get estimated distance, time and directions to reach a particular geo-location [13]. There is an administrator who has all Create, Read, Update and Delete (CRUD) privileges of the backend. The admin manages users and content. Lastly there is an Oauth Provider who is an external entity that provides authentication services. For this application the authentication provider in use is Google’s Ouath2 service. Figure 4 shows the use-case diagram of the system.

A data flow diagram represents the processes, data stores, and external entities in a business or other system and the connecting data flows [16]. Data flow diagram can also be described as a graphical depiction of the flow of information in an information system. There are different levels of data flow diagrams depending on details. The first level is called context-level diagram which shows the whole system as an overview and how data flows between the different actors. The next level is called level 1 and contains more details. Subsequent levels are called level 2, 3 etc. each with more details than the previous level. Figure 5 shows the context-level diagram of the CSO mobile application. Data flow from user to the system is authentication details for login or registration. In addition

user can upload their CSO details to the system. They can also browse and search CSO details from the system. Another data flow is from the administrator of the system. The admin has full CRUD permissions so they can send and retrieve data from the system.

The security of the system has also been taken into consideration. All communication between the frontend and backend is secured using Secure Hypertext Transfer Protocol (HTTPS). HTTPS is a protocol that is used to secure user data between a client and web server [17]. HTTPS is necessary to prevent phishing and man-in-the-middle attacks.

User passwords are stored in the database securely by using salting and hashing techniques. Passwords are not stored in plain text as this would compromise the system since any attacker who gains access to the database could read them. Instead passwords are salted in 10 rounds and hashed using bcrypt security nodejs module. Bcrypt is a scheme for hashing passwords. It is based on the Blowfish Block Cipher [18]. Bcrypt is resistant to brute-force attacks and greatly improves the security of a system [19].

User authentication is implemented using the JSON Web Token scheme (JWT). JWT is a standard used to create access tokens for an application [20]. The workflow of JWT begins with the client sending login or registration details to the REST API of the server. Once the server receives the username and password, it generates a token that certifies the user identity, and sends it to the client. The client stores this token and on subsequent requests, the client will append the token with the request back to the server. On receiving the request with the correct appended token, the server will know that the request comes from an authenticated client and not a fake client. In this application the JWT token is stored on the android mobile application using Jetpack Data Store [13]

Table 2 describes how the system requirements have been fulfilled by the designed system

Figure 6 shows the overall results of system evaluation. The system was evaluated by giving a number of users questionnaires with a set of questions and tabulating the frequencies of their answers. In the evaluation process a total of 60 users; whereby 48 users were from CSOs and community groups and 12 from marginalized communities answered a set of eight questions about the system. The results show that on average 88.125% of users granted positive reviews of the developed system in accordance to different criteria such as its ease of use (usability), accessibility, security, speed, availability and enhancement of information availability of CSOs. The remaining average of 11.875% represents users

who were unsure of some of the meaning of the concepts like security, availability and accessibility. Hence, the study received less numbers of users who were not satisfied with the system’s performance. In some of the issues clarification and training were carried out in order to assist the unsatisfied users with proper and clear explanation or information.

Software testing is a method of finding errors in an application or program. It ensures that the application functions according to the requirements [21]. There are three kinds of tests namely unit testing, system testing and integration testing. Unit testing is done at the unit level while integration testing tests the integrated units of a module. System testing is done to test the overall functionality of the system.

In the development of this application unit testing was done automatically to each unit by scripts generated by JUnit available on Android Studio [22].

User interfaces provide the first point of interaction between the users and the system. They are an important component of an application because they give the first impression of what the app can do to the user. Figure 7 and Figure 8 show some screenshots from the developed application.