FieldSight

Infrastructure is an enabling component for all seventeen Sustainable Development Goals. However, the implementation of local infrastructures often falls short of the promise they hold to transform lives. Poor-quality infrastructure occurs in the absence of quality assurance and construction monitoring that catches critical mistakes. FieldSight is the first tool of its kind to use mobile and web technology for quality assurance and monitoring of infrastructure projects. The FieldSight mobile app connects people and organizations working in remote field locations with engineers, designers, government agencies, and donors who can help provide guidance and oversight to help meet quality standards. Field teams who are embedded in local communities are able to collect data, receive advice, and contribute to a process that helps review, assess, analyze, and act on that data for the betterment of local infrastructure. In essence, FieldSight closes the loop to ensure that resilient designs become resilient communities.

Effective infrastructure management links processes across the infrastructure lifecycle, ensuring that decisions meet the evolving needs of the population. However, data is often siloed into disconnected processes that make effective management difficult. As a result, many public sector organizations cannot understand the state of their infrastructure then make plans for effective improvements. Post-disaster analyses after Nepal’s 2015 Gorkha earthquake revealed key weaknesses in oversight and construction that led to the lack of inclusion of seismic resistance, even where designs and local buildings called for it. Around 300,000 people perished due to inadequate buildings in Haiti’s 2010 earthquake but a much larger earthquake hit Chile six weeks later and the death toll was 600 times lower. Without reliable quality assurance data, local staff in charge of oversight may not have a productive way to deal with the pressures of corruption and any other cultural norms that marginalize them. Low capacity amongst local contractors and engineers make it difficult to construct higher-quality and more advanced infrastructure and conduct oversight. Improved capacity-building amongst local populations to inform resilient construction methods and enforce quality assurance inspection could help prevent tragedy and contribute to more equitable societies. 

Humanitarian and development agencies use FieldSight to collect more usable and actionable data to help populations directly affected by rapid and slow onset disasters. The mobile app helps to better include community-based field data collectors deployed by these agencies (end users like contractors and engineers) in the data collection process. According to research by the Global Alliance for Humanitarian Innovation, many humanitarian tech innovations fail because they are not created with the end-user in mind, but are rather created to meet the needs of report writers. The user interface for most data collection apps is built for high-level management to compile reports based on complex log models. The interface is therefore overly technical with monitoring and evaluation jargon that makes it more difficult for end users to operate. Misunderstandings during the field collection process are difficult to fix from afar which can make this process and later analysis phases less efficient and can even produce faulty reports. FieldSight is designed to ensure end users are informed about quality standards and how to monitor them. In this way, FieldSight also helps direct beneficiaries of infrastructure projects to be better represented in the projects concerning their communities.

FieldSight consists of mobile and web apps that create a closed-loop data collection and management system that help ensure that stakeholder perspectives are included in projects. This is especially important for stakeholders in the field whose voices can be misrepresented when data collection processes are not implemented well by headquarters-level offices. The mobile app is designed to support and enhance the work of those managing and monitoring projects in the field. They can use the mobile app to record information about the status and progress of infrastructure projects, access key information about the project and quality standards, and communicate with project managers. The web app allows technical and management staff to review data from the field, communicate with field staff, and explore and analyze data at the site, regional, project, and organizational levels.

The FieldSight platform builds on many recent advancements in the use of mobile technology and data aggregation and analysis while introducing a number of key features that enable the platform to more effectively support infrastructure projects and programs. 

• FieldSight includes features to assess and comment on field submissions, share comments between field users and headquarters offices, share site documents, and upload manuals and other guidance documents. This contributes to better, more effective, and real-time communication to ensure all voices are represented. 

• FieldSight ensures that all information can be accessed and collected in offline mode, ensuring that the app can be used in even the most remote or un-serviced locations. 

• Geo-locations and time stamps are built into all aspects of the system and can be used for oversight and verification of system users and stakeholders. 

• Through dashboards, reports, and maps, FieldSight enables the review and analysis of data at multiple levels, ensuring that lessons and insights from data can be seen not only site by site, but increasingly across projects, programs, and regions. 

As FieldSight scales, we plan to develop an infrastructure asset management platform that enables analysis by climate and population change to help field staff improve responsiveness to the needs of communities. Infrastructure status and management data collected via FieldSight throughout an asset’s life cycle will be used to analyze the ways population and climate affect asset performance. This data will contribute to an AI predictive model for asset performance that governments can use to better inform planning and decision-making. This will help make issues faced by communities more visible than ever.

FieldSight leverages the advantages of existing humanitarian project monitoring methods by coordinating them into an intuitive platform to help projects reach their full potential. Until now, monitoring and quality assurance of infrastructure projects have suffered from overreliance on qualified engineers and inefficient dispersed documentation techniques. Most oversight protocols include a trained engineer to provide regular oversight and guidance on site. Unfortunately, there is a shortage of trained engineers in many developing countries, especially in remote locations, so supervision is carried out by junior engineers, contractors, or community members without the requisite skills and knowledge. Most monitoring processes are documented on paper or via multiple third-party mobile applications such as WhatsApp. This is a cheaper method up front but occupies a great deal of time when the data needs to be organized and digitized for analysis. Moreover, it is impossible to identify key issues from individual sites and see patterns of challenges that affect multiple sites.

FieldSight offers offline data collection and secure data uploading for the infrastructure sector that has previously been unavailable. Kobo Toolbox survey software is built into the FieldSight platform so surveyors can digitize many types of data from the field, including photos and signatures. FieldSight enhances this technology by automatically organizing raw survey data into an easy-to-use platform where it can be contextualized for analysis and reporting. FieldSight’s mobile application includes features that help ensure high quality work, including offline educational materials and timestamps and geotags on all data submissions.

FieldSight’s web application is used to customize monitoring forms while the mobile app allows offline collection of data that can be synced later. The web backend of FieldSight is written in Python (Django) with a PostgreSQL Database. The mobile application is android native. 

FieldSight is built on two open-source technologies: KoboToolBox and Open Data Kit (ODK). KoboToolBox is a stack of technologies for custom digital form creation (kpi), submission receival (kobocat) and web data entry and preview (enketo express). ODK is a library that parses XML files into mobile forms and eases offline data collection by the use of the mobile device’s local database. Forms developed in KoboToolBox are downloaded and parsed in ODK allowing structured data collection, completely offline. 

FieldSight’s mobile application is built on ODK and has expanded functionalities to manage users and forms, two-way communication, notifications, and editing of past submissions. As an ODK-based app, FieldSight has offline data collection capacity. The FieldSight layer developed in Python (Django) supports projects and user management, form assignments, data export, dashboards and visualization, and more. FieldSight interacts with the above applications for the management features, exports generation, and visualization. 

Forms created and data received as responses are stored in XML format. Data is stored in MongoDB by KoboToolBox. PostgreSQL and Mongo DB thus serve as backup databases in the system. 

The databases and backend files made available through Amazon Web Services allow FieldSight to conduct large scale data collection and report users’ needs. 

FieldSight is a valuable tool because it can be used by field data collectors and infrastructure project manager without direct help from the FieldSight team. We also continue to improve the product based on our work with partner organizations. Our partners have inspired some of the most valuable features on the platform; such as creating site types, region designations, flagging forms, and the ability to map sites by progress level and other indicators.

Project success indicator: Infrastructure and other public projects are of higher quality because FieldSight’s features help teams better monitor and implement needed changes to projects.

• Sub-indicators of quality:

  • Timely project delivery that affects projected community benefits as little as possible

  • Partners are collecting correct data to support indicators

  • The product is built for the correct purpose and the way it was designed

• FieldSight Features that support this indicator:

  • Form Library: Provide globally reviewed and certified forms and form sharing capacity where users can share certified forms and indicators for any project type

  • Dashboards that support data management needs

• Evidence: Interviews with partners to understand how they use FieldSight (qualitative)

Product success indicators

• Increase # FieldSight auto signups

  • Means of verification: Internal records from FieldSight app

• Increase # of partners receiving 1:1 service

  • Means of verification: Cost Recovery records

• Increase # clients that scale up their data collection using FS (Assumption: FieldSight fits their data needs and they are using the product to scale)

  • Means of verification: Count of clients that use increasing tiers of FieldSight for their project(s)

The below numbers are based on the context of Nepal since our team has the most direct experience in this country. However, as we intend to scale globally, it is difficult to indicate a fully accurate number of beneficiaries.

• 200,000 houses at an average household size of 5 people/household= 1,000,000 people (4.6 people/household officially but rounded up for ease of understanding). (Government of Nepal National Planning Commission Secretariat Central Bureau of Statistics)

• 2,500 Schools at an average of 208 students per school (35,601 schools for 7,391,524 students)= 500,000 (Government of Nepal Ministry of Education, Science and Technology)

• 10,000 solar energy facilities with 100 people per solar energy facility= 1,000,000 people

• 2,500 health clinics at an average of 200 patients per health clinic= 500,000

• 10,000 km of Roads at 50 people/ km of road= 500,000

Long-term success will be measured by the extent to which infrastructure performs in alignment with the baseline standards determined during planning and design. However, as most infrastructure has life cycles of 20 years or longer, assessing this metric may not be viable in the context of this project. 

Proxy measurements for performance can be used to determine the impact of this project in the next year. The timing of the first repair of infrastructure and the frequency of repair are key proxies for overall performance. Infrastructure performance in the face of population growth and disasters, both more frequent events, can also serve as proxies. 

Finally, while not addressing the overall impact, the value of the system for decision-makers can greatly influence its uptake. User metrics and user surveys will be analyzed over the next 1-5 years to determine the extent to which governments and key actors find the system useful.

• Risk 1: High risk, low probability: FieldSight provides a false sense of security. If partners have poor monitoring standards or give FieldSight to field staff without the proper training, the projects could still produce poor quality data and outcomes. 

• Risk 2: Medium risk, low probability: Ongoing costs to using the system exceed available budgets of users. 

• Risk 3: Medium risk, medium probability: There is an incompatibility between the FieldSight system structure and monitoring standards of user organizations. 

• Risk 1: To mitigate this, we would use our training materials to stress the importance of developing forms that encourage the collection of robust data and high standards for training field staff. Where possible, work with partners to develop this structure. Share “certified” forms that have been approved by known experts. 

• Risk 2: To mitigate this, we plan to expand the user base to spread out core management and development costs. Itemize costs for the system, set-up, training, and other needs. so organizations can work with FieldSight within their budget. Our open-source license allows skilled organizations to adopt FieldSight independently at no extra cost.

• Risk 3: To mitigate this, we would work with projects and partners early on to develop standards in a way that matches FieldSight’s system structure. Develop certified standardized forms that are compatible with FieldSight and available to all partners.

United Nations Office for Project Services: FieldSight is a cost-recovery program housed inside UNOPS. We also work closely with a development partner, NAXA Ltd., a Nepali company that helps develop and distribute the software. UNOPS provides the institutional home and logistical support to work in more remote parts of the world, while the FieldSight team and Naxa are responsbile for strategy, management, administration, and finances, including generating revenue to cover costs.

20 full-time staff

Our team comprises diverse skill sets and experiences in contexts familiar to our clients in the humanitarian sector. Our staff includes data scientists, a designer, web developers, communications specialists, and experienced mobile survey creators. Furthermore, our team has several years of experience helping clients in the infrastructure sector and beyond to apply FieldSight to their monitoring needs.

• Naxa (web development partner)

• The United Kingdom Department for International Development (DFID)

• US Agency for International Development (USAID)

• UNOPS Asia Region

• UNFPA Nepal

• People in Need

• Government of Nepal National Reconstruction Authority

• World Vision International Nepal

• IFRC Nepal

FieldSight is an award-winning easy-to-use platform for improving quality and reducing risk in infrastructure development. Our key customers are entities that fund most infrastructure development: governments, international finance institutions, humanitarian agencies, NGOs, and private companies. While other competing platforms exist, they do not target the project modalities unique to the infrastructure sector. Also, competitor mobile data collection software, such as Kobo Toolbox, is not built to support process management that links activities, outputs, and outcomes. 

FieldSight’s main revenue sources:

• Access to the platform: FieldSight is open source and can be programmed for free but the majority of users choose to subscribe to the platform by signing up themselves via our website or contacting our team. These users can be individuals or teams and often introduce others to the platform. They receive the entire platform ready out-of-the-box to be customized to their project. We also aim to build institutional partnerships where organizations use FieldSight across their entire portfolio; we believe this would standardize monitoring practices for large organizations and would also stabilize our user base. 

• Support services: Many of our users in the infrastructure sector are new to mobile monitoring techniques and using FieldSight helps them evaluate their overall quality assurance process. Part of our revenue stream includes training and set-up services for users and we are always improving these trainings. Furthermore, we can also help develop monitoring systems for organizations that want advice on ways to structure monitoring processes and systems to ensure infrastructure quality and resilience.

We aim to continue using FieldSight across all UNOPS operations in Nepal, scale the application to all Asia Region UNOPS projects (agreement in-process), then scale to application globally in UNOPS. This will also provide more opportunities to connect with global funders and humanitarian actors. At the same time, we want to broaden the use of FieldSight beyond the humanitarian community. Starting in Nepal, we will work to add more government and private sector users and build a model for future partnerships. We will then aim to expand to additional countries, starting with neighboring South Asian countries, then identifying additional potential markets elsewhere. Scaling up within UNOPS and the humanitarian community will rely heavily on our partnership with and colleagues at UNOPS; however, as we expand private sector and government use, we aim to add more business development, government liaisons, and support staff directly within our team. 

While FieldSight has scaled up in UNOPS through more than 16 countries, we have several goals that could be achieved with assistance from the Solve grant and the community of Solvers. FieldSight is a tool that needs to be used with best practice monitoring standards so our team needs to find ways to help our partners use FieldSight responsibly.

We want to create create community portals where infrastructure users or beneficiaries can give feedback through an app or website. We would then provide data to these communities for review and approval. This feedback would triangulate with data being generated by our other institutional partner organizations that use FieldSight to monitor their projects. In this way, we can improve feedback to communities beyond the end-users who conduct surveys.

Using artificial intelligence, we want to use FieldSight data to identify common problems within a project so they can be solved more rapidly. While AI cannot be the only solution to improving monitoring standards, it can help users apply high standards more consistently.

Solve would provide an important network of like-minded professionals with whom we can collaborate to find more grassroots issues to tackle. FieldSight is designed for end-users in the field and we want to continue to work with Solvers that can help us better work with our users and anticipate more of their needs.

In the infrastructure space, we would like to partner with public and private sector partners, and eventually scale FieldSight to be used institutionally across government entities as well. We are seeking institutional partnerships with entities like UNOPS, the World Bank, the Asian Infrastructure Investment Bank, World Vision, Oxfam International, and ministries of public works in national governments. These entities control many projects globally where FieldSight can make an impact. 

Adding AI features to FieldSight would help our users collect data in more uniform ways, share it more widely, and analyze it more effectively. First, we will develop a form library of certified robust forms to help users collect better data. Additional features to flag and annotate responses and photos will then allow more granular data analysis. Second, we will strengthen data-sharing features to help organizations create custom reports, share data in public portals and directly with partners, and link data to other systems. Data security would be paramount so privileged or sensitive information does not need to be shared. Third, we plan to introduce machine learning by developing algorithms that review users’ flags and annotations then use that data to predict incoming submissions that might be flagged or in need of review. This will be especially valuable for large-scale post-disaster projects. 

Our AI model algorithm will be built on ongoing user input throughout the lifecycle of their infrastructure project. Users will provide data to an AI model by flagging questions and issues in the construction process and reporting on the accuracy of FieldSight suggestions. Our development team will simultaneously review accuracy rates and adjust the model if accuracy rates drop. In order for this AI initiative to be most effective, we need to have a consistent user base so we have enough data and resources to continue developing this and other features.