Citizen-driven Household Dietary Diversity Dynamics (Citizen-H2D3) for food security
Citizen-H2D3 is a data-rich decision-support system that empowers citizens to improve dietary diversity & national food security in Rwanda.
This video provides a brief representation of Citizen-H2D3 as a data-rich decision-support system that empowers citizens to improve dietary diversity & national food security in Africa (Pilot in Rwanda). The system will leverage on citizen engagement to provide evidence-driven and robust insights about the dynamics of food (and nutrition) security as an outcome of food consumption and dietary diversity while unraveling linkages with environmental and eco-social covariates in space and time.
Lead Applicant Organization Name
International Institute of Tropical Agriculture (IITA)
Lead Applicant Organization Type
Large NGO (over 50 employees)
If part of a multi-stakeholder entity (i.e. team), provide the names of other organizations and types of stakeholders collaborating with you.
1. Rwanda Agricultural and Animal Resources Board - RAB (Government, Rwanda)
2. Bioversity International (Large NGO)
3. University of California, Davis (Other)
4. Viamo, Rwanda (Large Company)
5. European Commission Joint Research Center (Other)
Website of Legally Registered Entity
How long have you / your team been working on this Vision?
Lead Applicant: In what city or town are you located?
Lead Applicant: In what country are you located?
Your Selected Place: what’s the name of the Place you’re developing a Vision for?
Rwanda, a small Country in East and Central Africa, with a total land area of less than 27,000 km^2.
What country is your selected Place located in?
Describe your relationship to the place you’ve selected.
Rwanda is a country that has shown great potential to leverage digital technologies for development and considered as a prime Country in sub-Saharan Africa to pilot the innovative Citizen-H2D3 system. I relocated to Kigali in July 2018 to lead a GIZ-funded project. The project is focused on developing digital surveillance tools to combat a major banana disease, which invariably constitutes a threat to national food security in the Country. As part of the project, I led a baseline survey of ~800 households across ~140 villages and 8 districts across the Country. Our team’s interactions and insights from data provided compelling evidence regarding gaps in food and nutrition security which varied with locations and with time (mainly seasonally). Although various programs are being implemented by the government to reduce inequality & improve access to crop production inputs, there is an extant need to accelerate the reach and effectively direct resources/interventions to where they are most needed. Also, many reports have documented issues of malnutrition and household food insecurity. For instance, USAID (2018) reported that 1 out of every 3 Rwandese experience food insecurity, while the Country suffers from the double burden of malnutrition, with 21% of Women suffering from overweight and obesity. The problem of food and nutrition insecurity is not merely prevalent in rural/remote areas of the country, but this major problem spans across the rural-urban continuum, typically due to lack of actionable decision-support/advisory for individuals to make the right dietary choices, or for relevant institutions to guide their programmatic priorities on food security. Engaging citizens to generate and utilize such rich-data over time is indispensable for viable actions and equitable access to nutrition needs within Rwanda’s food system.I have been impressed by the opportunity to engage citizens and policy actors in Rwanda with promising digital innovations and strategies.
Describe the People and Place: Provide information that would be helpful for an outsider who has never been there and may have no context about this Place to better understand the area.
Rwanda is located in the heart of Central and East Africa, with easy access to neighboring countries (including Uganda, Tanzania, Burundi, and Democratic Republic of Congo). It is often called “The land of a Thousand Hills" due to the predominantly green undulating landscape of hills, gardens and tea plantations across the Country’s 26,338 Sq km land area. Since the country lies south of the Equator and at relatively high altitude, Rwanda is characterized by a tropical highland climate. Temperature varies considerably between locations, but the monthly averages per location do not vary significantly. Peak temperatures rarely exceed 30°C (86°F) in the daytime or drop below 15°C (59°F) at night, while the average daily temperature in the capital city (Kigali) hovers around 21°C (70°F).
Generally, the land area stretches across elevation ranges of 1,500m and 2,500m, with the peak (Mount Karisimbi) at 4,507m and the lowest point (Rusizi river) at 950m above sea level. The official languages in Rwanda are Kinyarwanda, English, and French, however majority (>90% speak Kinyarwanda). Its capital city, Kigali, is frequently ranked the cleanest city in the region. The country has the highest population density in Sub-Saharan Africa, with estimated population of 13 million people. About 62 percent of the population lives on less than US$1.25 per day (USAID 2017), yet, Rwanda is a thriving, safe country with one of the lowest crime rates in Africa, despite the prevailing low-income level.
Agriculture is the mainstay of the Rwandan economy because it generates 29.5% of GDP, provides 45.9% of employment, accounts for 63% of foreign exchange earnings, and provides 90 % of the country’s food needs. However, crop yields are considerably low due to various factors. Majority of farmers practice subsistence farming on small, hilly plots which are often subjected to erosion and poor soil fertility practices. Over 50% of the country’s farmland is severely degraded, and the use of improved seeds, pesticides, and fertilizers is still limited (USAID 2017). Since the launch of crop intensification program in 2007, increased production of food and cash crops have been reported. The crop intensification program aims to increase agricultural productivity in high potential food crops namely maize, potatoes, wheat, beans, peas, cassava and soya. Recently, the attention of the national government (through Rwandan Agricultural and Animal Resources Board - RAB) is shifting to vegetables as a critical component for proper nutrition and diets.
Rwandan culture is somewhat woven around the unique music and dance which often featured in ceremonies, social gatherings, festivals and storytelling. The celebratory dances are usually accompanied by an ‘orchestra’ of drums and nine energetic men who enthusiastically provide the beat. ‘Intore’ as it is famously known is the most famous traditional dance and it consists of highly choreographed dances.
Map showing location of Rwanda in East and Central Africa.
Zoomed-in aerial view at Kanombe, in Rwanda showing density of households.
One of the landscapes that are typical in Rwanda, earning it the appellation "Land of thousand hills"
A digitally-engaged buyer in the market where food commodities are sold in Rwanda (Picture taken at Kigali, Rwanda's capital City)
A typical market where food is sold in Rwanda (Picture taken at Kigali, Rwanda's capital City)
What is the estimated population (current 2020) in your Place?
Challenges: Describe the current (2020) and the future (2050) challenges that your food system faces.
Rwanda is a small, landlocked country with the highest population density in Sub-Saharan Africa, and it is one of the world’s poorest countries. Considering that 2 out of every 3 Rwandan population lives on less than US$1.25 per day (USAID 2017), the challenge of feeding the growing population remains daunting. Generally, malnutrition during childhood and pregnancy is adversely consequential for child survival and long-term wellness. Further, it poses a significant threat to human capital, economic productivity, and overall national development. The inevitable impacts of malnutrition (if unaddressed) is a major concern in Rwanda, where 38% or 661,200 children below 5 years suffer from severe malnutrition (stunted height) and 37% or 643,800 suffer from anemia, based on recent data from Demographic and Health Survey (DHS) conducted by the National Institute of Statistics of Rwanda [NISR], Rwanda’s Ministry of Health [MOH], and ICF International (USAID, 2018). In the future (post-2020 to 2050), we expect the emergence of a different shade of the malnutrition challenge. As the population grows and more citizens move to the major cities, there may be an increase in obesity and overweight cases due to access to unhealthy foods that meet the fast-paced lifestyle of the city but offer limited nutritional quality to support good health.
There are 3 major dimensions to this challenge: (i) the lack of information about dietary deficiencies, (ii) the low affordability, and (iii) the gap in institutional targeting of intervention due to the unavailability of actionable data/information. The first 2 dimensions invariably pertain to individual Citizens, while the 3rd dimension accrues to governments and stakeholders in the food system. Since the end of civil war and genocide in 1994, the Government of Rwanda has embarked on rebuilding the country and improving the quality of life for its people. Yet, the current income and food security status indicate that major gaps are persisting in the food system both in terms of quantity and quality of the citizen’s nutrition and diets. Current data on food/nutrition insecurity are limited to annual/biennial surveys, with limited spatial and temporal disaggregation. Addressing this dynamic problem, with only static approaches can only provide very limited snapshot insights to this exceedingly complex problem, and may constrain the proper deployment of intervention and resources within the Country.
Address the Challenges: Describe how your Vision will address the challenges described in the previous question.
We envision that Citizen-H2D3 will empower citizens for dietary decision-support and unravel the spatio-temporal dynamics of food and nutrition insecurity in Rwanda by generating robust data-driven insight that - i) leverages on easy-to-deploy mobile-based tool to collect spontaneous and rich data directly from citizens; ii) provides [near]real-time intelligence on household dietary diversity to engage relevant stakeholders and development partners; and iii) links rich georeferenced data on food and nutrition to biophysical, climatic, and ecosocial data, over space and time, as a decision-support mechanism for interventions and investments on malnutrition. This will provide a rich learning platform for understanding the direct and lagged interactions between extrinsic and intrinsic factors that underlie rural-urban food and nutrition security.
At the piloting stage, the Citizen-H2D3 system will be uniquely set-up to motivate (without compelling) the participating VCs to submit data on a daily/weekly basis, yet imbued with sufficient controls to allow for non-dominant participation of each entity over time (based on the incentive structure). Volunteers will capture and submit dietary consumption data, via Open Data Kit (ODK) forms, including type, source, and amount of food group that they (or their household) consumed for the specific day of data reportage. Each volunteer will participate by downloading admin-formatted ODK forms to their smartphones. Citizen-H2D3 will move away from traditional utilisation of periodic surveys and recall diaries (which are often limited in capturing the long-term nuances of household food consumption), and will track a series of metrics (dietary diversity, food frequency, household food consumption (diversity * frequency), and food source diversity) in [near-]real-time. Tracking these metrics across space, time, and income brackets, will parse the dynamics of household food and nutrition security Rwanda, and position stakeholders within the food system to deploy contextually-relevant strategies to improve quality and facilitate equitable access of proper nutrition by citizens.
High Level Vision: With these challenges addressed, now provide a high level description of how the Place and the lives of its People will be different than they are now.
In the short-term, we envision that Citizens will be empowered to “know” and “act” on their dietary health/status, especially within the urban and peri-urban areas (and later the rural areas). At the front-end of the system (the mobile-based app), the volunteers/users will receive quantitative feedback (metrics) about their dietary health/summary on a weekly basis, with a suggestion of locally available food options to improve their dietary diversity and nutritional status. By empowering citizens with the decision-support/decision-advisory tool, we hope to elevate the “individual-awareness” of adopters and their households (in Rwanda) about balanced nutrition. In the first 2 years, we expect that the dietary diversity index of volunteer citizens will have increased by ~50%, relative to baseline, with the adoption of the tool by at least 10,000 Rwandese.
In the long-term, the periodic (daily) data flow from individual citizens/households, across the country, will be consolidated in space and time, fed into data-crunching analytical algorithms, and used to generate insights through open-access interactive visualization dashboard (the back-end). For instance, extensive analyses of data from various household types, income, locations, and ages will provide relevant insights than can support institutions (government, NGOs, development agencies,and strategic thinktanks) to prioritize when and where to target intervention-driven resources to tackle chronic malnutrition. Based on the democratized structure of Citizen-H2D3, women and youth of all age groups can access, contribute, & benefit from this initiative.The system will bypass the dominant patriarchal culture of the society because it will be implemented on smart mobile-based technology, which is often personal and accessible to the owners. We anticipate that the final disaggregation of the data by gender and age will reveal compelling data-rich evidence to inform gender-balanced solutions to malnutrion.
Full Vision: How do you describe your Vision for a regenerative and nourishing food future for your Place and People for 2050?
Food and nutrition (in)security are both global issues. The number of undernourished people reached 817 million in 2017, with the problem being more chronic in East Africa. A shift towards nutrition-sensitive food systems is needed to promote and provide access to healthy diets. Despite this need, limited data and monitoring systems exist to assess the dynamics of this problem at scale. Generally (when available), short-term, fragmented, and static data on food systems poses a significant drawback for the formulation of location-based and time-relevant policies or action to ameliorate food/nutrition insecurity nationally or at household level. Our visionary idea is to develop a [near]real-time system for monitoring and improving dietary diversity in space and time, based on a citizen-driven spontaneous crowdsourcing approach in Rwanda (where food and nutrition insecurity is a major problem). In addition to empowering individual citizens, a nuanced and disaggregated insight into food consumption across space and time will be valuable to relevant institutions and actors in the Country’s food system to guide policy strategies and interventions.
Our visionary idea (Citizen-H2D3) will leverage on a broad collaboration to co-develop and co-validate an innovative ICT-driven approach for inclusive advancement of the national food system in Rwanda, in alignment with the government’s strategic priority. Volunteer citizens (initial-cohort/beta-users) will be invited by accessing contact databases (such as the 845 USSD service platform and Smart Nkuganire Farmer subsidy platform) and through multiple media and print campaigns. At the core, we will adopt a participatory human-centered design (HCD) approach (Steen, 2011), with focus on the representativeness of rural-urban Rwanda, to develop an interactive system (front-end smart tool + back-end analytical database) that empowers the citizens to track and receive recommendations on their dietary health.
Citizen-H2D3 will be initially piloted as a multi-platform system, including the use of free android-based Open Data Kit (ODK) App and cloud-based server, to co-develop and tweak workflows with the beta-users (initial VC cohorts). As the system matures, it will evolve into an end-to-end interactive and stand-alone system, with availability to public users who can sign-up to use the tool, receive feedback on their household dietary diversity, and access recommendations to improve their dietary decisions. This engagement is expected to provide intrinsic nudge for users, and Citizen-H2D3 will evolve under as an information-based incentive model.
Our progress will be step-wisely assessed based on the two (2) critical benchmarks of this idea. The first benchmark will be focused on successful co-development and co-deployment of tool for efficient dataflow and algorithm-driven recommendations, while the second benchmark will be focused on system performance. The first benchmark will include 4 phases of progress monitoring during the initial 24 months (M1 - M24), including - Phase 1 (M1-M4): Initial system set-up completed (including co-development of relevant standard operation protocols (SOPs), launch of information-rich project googlesite, collaboratively composed and tested ODK forms, vetted data management plans, micro-reward system set-up, and public call/campaign for prospective volunteer citizens; Phase 2 (M4-M8): On-boarding of verified volunteers completed with anonymous ID assigned and admin support deployed in readiness for data submission, and visualization platform developed for beta testing; Phase 3 (M8-M20): Data-flow steadily maintained and micro-rewards periodically dispatched to qualified volunteers, and interactive visualization platform validated with initial datasets; Phase 4 (M20-M24): Space-time analyses completed with the linkage of indices to eco-social and biophysical/environmental variables. Further scenario modeling (ca. 2050) of food and nutrition security in the Country will also be implemented based on initial relationships with the ancillary variables and considering strategic policy options.
The second benchmark will focus on the overall technical performance of the Citizen-H2D3 system. This will be assessed based on 3 criteria namely: i) volunteer crowd performance; ii) data quality; and iii) cost of implementation. The progress on crowd performance will be assessed against our targets, which include the on-boarding of 1,500 volunteer crowd members within the first 8 weeks and consistent submission rates or data flow (300/week and 45/day data records); Data quality progress will be based on consistency of (low) rates of erroneous data points (i.e. <5%) based on completeness and correctness checks; and cost of implementation will be based on calculated cost of acquiring each datapoint. All these will be assessed based on indices which will be calculated on a weekly basis. For instance, the crowd engagement index (CEI) will be calculated as the ratio of volunteers that submitted data in a week compared to the number of this enlisted. The minimum value will be preset at 0.1 (i.e. 10%) per week.
Prospective volunteers will be offered instructions to sign-up and follow rules for participation in the co-development process. A minimum of 1,500 volunteer citizens (VC) should be onboarded before initiation (within the first 4 weeks). We expect consistent submission rates of 300/week and 45/day data records due the financial-based incentive system. After submission of profile data, VCs will be assigned an ID (VC_ID) and they will be required to follow basic guidelines (on project’s googlesite) to install the ODK-app on their smartphone and complete the ODK-deployed e-survey at will. The content of the e-survey will include 10 food groups and VCs will select dietary make-up over the past 24 hours. For each checked food group, volunteers will be presented three subsequent questions: i) amount eaten (e.g. plateful, handful, cupful); ii) household members who consumed; and iii) sourced of the food (e.g. produced, bought, gifted). This filled form will be sent to a central data repository and automatically recorded for the VC to be considered for the incentive.
The incentive will be designed as a gamified mechanism to keep the VCs consistently engaged over time and sufficiently inhibit/minimize risk of spamming. Meager, yet non-promissory, incentive will be randomly offered to 10 (out of >45) VCs that submit data daily. VCs will be instructed to submit throughout the course of the day, but submitted data should be based on consumption from the previous day. To maximize system fairness and possibly increase the diversity of contributions, maximum threshold of reward for each VC will be capped (1/week and 4/month). This approach is expected to effectively compel each VC to competitively submit data in a timely manner (at least weekly), and equally dissuade them from excessive/redundant submissions. We assume ~20% (i.e. 300) of onboard VCs (~1,500) will submit data, at least once per week. To ensure quality of data, random follow-up phone-calls will be made to 2-5 participants daily, validating the consistency between e-survey response and verbally retrieved responses. To avoid ‘hacking’ of the H2D3 incentive system, only the first submission will be recorded per VC_ID and only one submission will be recorded per geo-referenced household (precision of 3m) using geospatial buffer analysis.
We reckon that the sustainability of Citizen-H2D3 will be based on long-term access of citizens to smartphones. Although mobile phone penetration has increased dramatically (6-67%) from 2006-2017, smartphone penetration remains low. A price point of (entry level smartphones) ~$20 is likely to encourage rapid penetration in Rwanda. Initially, limited smartphone access may reduce the representativeness of data captured, with information skewed towards younger, urban, and citizen who are more educated/professional. However, Citizen-H2D3 is proposed as a scalable, long-term, and efficient data collection system, which will become increasingly relevant with the wider penetration of smartphones into Rwanda and sub-Saharan Africa.
Citizen-H2D3 will empower individuals and also generate nationally relevant insights about food and nutrition insecurity and where they are most severe in space and time to guide policies, interventions, and household level decisions on diets.
We aspire to empower women and children by giving them access to the decision support that they need, and giving proper visibility to their nutritional challenges to generate relevant multi-stakeholder actions.
Our logo and overall vision for Citizen-H2D3
How did you hear about the Food System Vision Prize?