You are here

What kind of data is most relevant for rural households? Is this data already being collected?

26 posts / 0 new
Last post
What kind of data is most relevant for rural households? Is this data already being collected?

In this part of the dialogue we will focus on what data can be shared with households rather than what data can be collected solely for research purposes.  As mentioned in our overview, we discussed the use of M-farms to provide market information in Kenya and Agro-hub to provide this information in Cameroon.  We also discussed the provision of weather information in a number of countries through M-farms and the provision on production advice through Plantwise.  Any experience on the use of the platforms, the information provided, and other information needs at a household level will be helpful.

Relevant data

Data that is most relevant for rural households will depend from place to place and to a large extent on the level of integration of rural households in markets.

Data needs

Data needs will be dependent on the location.  In terms of the types of data we might consider, there may be some commonalities such as weather or crop production information or local prices in neighboring markets.

What is Most Valuable?

We all have a bit of a data obsession at the moment! the BIG DATA movement is exploding into the Agriculture space and if you are living in the world of High Value Agribusiness and Precision farming, you need to be emmersed in the data lake.

BUT, for most smallholder farmers in SSA, they are still NET buyers.... i.e., they are pre-commerical, 70% are unbanked, their farms are untiltled, they rely on rainfed farming systems and very few get access to improved seed and fertilizer....    

In the information space, the data community (Governments, research, agribusiness, farmer coops, NGO's etc) the data that is being collected, the data that is being shared and the needs of the farmers are not well aligned..  

Andy Jarvis of the CGIAR's BIG DATA team, in his BIG DATA launch at the ICT4D  held last week in Hyderabad, said the following...  Agriculture in development still has a long way to go with our data opportunities. We have a lot of data, but its fragmented, not standardised, hidden, we are probably protecting a lot of poor data and make progress, we need to start sharing our data, (in a secure way!) so that we start to see whats of value to who? 
The data that a seed business wants is different from what a trader needs. Everyone has an idea what farmers need...  but from many conversations i have had recently, millions of farmers still need an intermediarly of some form, to help them navigate data / or information and i am not sure we have provided sufficient support to those intermediaries to be in a position to unlock the value of data to farmers yet...    

In echoing Shaun's point

In echoing Shaun's point about the need for intermediaries, I'll share some preliminary findings from one of our current projects. The focus was on sharing information about drought tolerant maize varieties to farmers in a drought-prone region of south central Kenya. Baseline surveys showed that there was little awareness of the newer maize varieties. Our intervention involved bringing this information out to rural communities, using a video created in the local language and featuring local actors and settings.  It was screened in churches and schools and the screenings were open to anyone in the community. We learned a lot, but just to skip to the intermediary point - we found that our interventions had much more impact in communities where there were village-based intermediaries (in our case - village based advisors who were part of FIPS Africa, an organization that helps provide access to farming inputs through a network of VBAs). Just learning about the new types of seeds and how to use them was not enough, there needed to be an actual source for obtaining the seeds (not every rural village has easy access to an Agrovet) and someone local to reinforce the messaging in the video. In those situations, the intervention worked, over and above having a VBA alone. So complementary strategies, involving use of new approaches to disseminating highly localized information combined with on-the-ground support from local intermediaries, seems like a good direction to go.


data needs of producers

I agree that more open data is needed.  However, as you note, it is important to be clear on what will be valuable to share.  I believe there should be more focus on collecting and sharing data that would be useful for producers to increase resilience to production or market shocks. There is a need for what this entails and the mechanisms for funding, sharing, and training on the use of this data at the local level.

Data Needs

In terms of the data needs another important information is the post-harvest support or services available. In Rwanda the government is investing heavily on post-harvest handling and thanks to the strong farmer cooperative movements this information gets to the farmers easily

Another area is on opportunities for value addition which I would link to agribusiness - Again there are some initiatives going on within the farmer cooperatives towards this

post-harvest support

Thanks, Mercyline.  I agree that post-harvest information is critical in many places.  What type of post-harvest support is being provided in Rwanda and is this funded solely by the public sector or through partnerships?

Post Harvest Support

There are community based post-handling facilities owned by farmer cooperatives.

They get support in perms of trainingson post-harvest handling technologies as well as marketing

The funding is through collaboration with donors


post-harvest technologies

Is this only done at the facility or is information shared via mobile or other technology? 

Post Harvest Technologies

I am not aware of any mobile platform sharing the information

Eduardo Nakasone
Post Harvest Technologies

Hi Mercyline, 

I think there is an increasing effort to produce  audio visual materials that can be reused and massively broadcasted (not only for agriculture, but also applications related to health, education, female empowerment, etc.). 

In the particular case of post harvest losses,  SAWBO (Scientific Animations without Borders) has created some videos that can be used to teach farmers effective techniques to reduce food losses: 

These videos are available publicly on Youtube. For example, this one teaches how to build a solar grain dryer with simple materials:






data needs

The small scale farmer data needs may be very simple like what are the new and best tomatoe variety for my area? whatsis there charcteristics? where can get seedlings, the weather data? Farm gate and market prices?

data needs

Thanks and I agree on the importance of relatively easy to use information.  Given your work with Plantwise, how has the team approached this (what data is provided, who is funding that initiative and for how long) and what data was most relevant for producers in terms of actual use?

Stepwise data

Totally agree with this...  There is a sense of order in the data option..  and where to start is often an issue?
Basic for the field agent? Farmer registration, Farmer location, Size and # of Plots, Key products...  (its a start? and its not being done really..)

Basic for the Farmer? (some assumptions..  a farmer who knows their market requirements.. would really be helped by ​Best variety, who sells it, When to plant and any key issues related to that product? I believe that the horticulture industry is now selling working with projects and Telco's to sell horticulture seed packs with a short code printed onto the package that enables, interested farmers, to access expert advise on that crop, via SMS.... 

this may all be a little clunky at the moment, but its a start in using technology to join some dots in terms of better genetics, advise and phones.... 




These are all great points,

These are all great points, there is not a one-size-fits-all set of data, especially if you go beyond the farmer and the field agent and move into data which is relevant for the ministry and regional organizations. 

Berber Kramer
Getting data by collecting it

One thing to consider as well, is that farmers actually have a lot of data already themselves, and that helping them to organize this data could be very valuable. I am thinking here for instance of demo plots where farmers are encouraged to apply new technologies/practices next to their existing technologies/practices to see the result. In an example that I experienced myself (although in India, not in SSA), farmers who were asked to regularly take pictures of their plots said in Focus Group Discussions that this improved their management practices because they had a better way of monitoring what was happening in the fields. In other words, directing farmers' attention to data that they might already have, or data that they are already collecting themselves, could also help address farmers' data needs.

Analysing data

Berber, you raise two great points..  What are collection and how can local service providers (public / private sector) help them to analyse their data..  I have the same feeling on this about several other players who have their own data..  In the day of data obsession, whether its a farmer, farmer organization, NGO, or private sector, there is data..  and many of these organizations are investing in ways of collecting data... BUT i get the impression that we and these different actors are not investing in analyzing data..  i think that we might find that collection LESS data, that helps a person / organization / company to make a decision is as valuable as having the capcity to collect lots of numbers?    

Use of existing data

I agree and I think it is important to focus on data that will actually be used by not only policy makers but also producers.  As Berber mentions, collecting some data from producers and providing information back to them is relatively cheap but to ensure usefulness and quality, it will be critical to support this from the public and private sector, adapted to local conditions and needs. 

Analysing data

Berber, you raise two great points..  What are collection and how can local service providers (public / private sector) help them to analyse their data..  I have the same feeling on this about several other players who have their own data..  In the day of data obsession, whether its a farmer, farmer organization, NGO, or private sector, there is data..  and many of these organizations are investing in ways of collecting data... BUT i get the impression that we and these different actors are not investing in analyzing data..  i think that we might find that collection LESS data, that helps a person / organization / company to make a decision is as valuable as having the capcity to collect lots of numbers?    

Data needs to further remote sensing capabilities

From the remote sensing perspective, there are a number of unique challenges presenting for agricultural monitoring in SSA. Field size and locations in particular place severe limitations on the usefulness of the systems which have been used in the past. For example, the MODIS sensors provide 250 meter spatial resolution with approximately a 2 day revisit time. This has proven to be too coarse for observing smaller fields, so we must turn to higher resolution sensors such as Landsat 8 and Sentinel 2, capable of imaging at up to 30 meter and 10 meter resolution respectively. The combination of Landsat and Sentinel is capable of providing higher spatial resolution with a short revisit time of approximately 5 days, which is a remarkable advance in the state of the science.

However, there are still outstanding issues - first, the data acquisition is still ramping up and we do not yet have global 5-day coverage; second, the volume of data is much greater and requires significant resources to process; third, because the constellation is relatively new, the data archive is short (compared to ~17 years of MODIS observations); and finally the existing models are not caliibrated for the higher resolution data in SSA. The last point is key - while the earth observations data for monitoring agriculture in SSA is starting to come online, high quality ground observations are needed to link the satellite observations with ground conditions. Well-distributed (both spatially and in terms of availability) data on biophysical variables such as crop growth cycles, crop height, yield, etc. can be used as input into monitoring systems which provide benefit for all.

Additionally, in many parts of Africa plant growth is limited by water availability. This makes datasets which monitor rainfall such as the Climate Hazards Group Infrared Preciptation with Station (CHIRPS) data essential. However, many of these datasets supplement satellite observations with ground station data - unfortunately the density of rain gauge stations is decreasing dramatically across the globe as older systems begin to fail. Installing new stations is absolutely necessary to ensure the quality of this very important data.

In the context of this topic, the data is collected for research, but in practice the derived data are made freely and openly available to anyone which in turn can theoretically benefit the individual household given strong means for communication.

remote sensing

How do you envision the data you mention being used by policymakers and producers to make decisions and what have been the limitations to doing so (for example where we have seen substantial drought and resulting famine)?  

From your perspective, how can the public and private sector not only support the data collection you mention but also dissemination in way that is useful to policy makers as well as producers?

Good policy should be founded

Good policy should be founded in good data. In order to know the effects of drought on agricultural production, it is necessary to know the extent of the drought and the location of the crops. But, even basic datasets such as crop location can be very difficult to develop and require a lot of time and expertise to create and maintain, particularly in regions where field size is small and crop rotations vary. So far, this has been the limitation for remote sensing which is starting to be addressed with new sensors.

The next challenge, as you are getting at, is communicating the information with policymakers. In Karamoja, Uganda, food aid was dispatched in response to a report which used remotely sensed data as justification based on in-season monitoring. However, this is not always the case and it takes time for people to trust in and adopt data sources. The approach at the national scale has been to involve the ministries of agriculture in the monitoring process by providing training and tools for monthly monitoring, but keeping the responsibility of creating and publishing reports with the ministry. The ministry independently draws its own conclusions and has ownership over the process. I think the concept of ownership over the data and process as a whole are very important to ensuring the process continues.

In terms of data collection, dissemination is a very important topic. From a producer's perspective, individual data points of the types I mentioned in the previous post are likely not very interesting. However, if that data can be used to in turn provide information on, for example, ideal planting conditions, there is a feedback loop which benefits both sides. The bi-directional flow of data may be one way to increase engagement.

bi-directional data flow

You raise a number of important points here. One in particular that has been noted by others, that perhaps more of a focus on a data feedback loop could help better ensure sustainability and usefulness.  There will need to be multi-stakeholder design and support of these types of initiatives but they may be more effective in the longer-term.  

how is Geospatial data being used

Certainly there is a great deal of excitment around the potential use of Geo-Spatial data and yet for most actors its still being used more for location than other forms, but there are signs now that more scaled uses are being used..

Plot area, there are several applications such as Fields Area Measure, which are using mobile phones as a means to help agents / farmers determine a more accurate knowledge of farm plot size. Area is probably the largest source of error in calculations on yield and profitabilty at the moment. So a really helpful thing to get right.

Imagery: The use of satellite imagery is also helping some farmers and communities, where this information is made available to be able to locate their farms and hold conversations around borders, boundaries and potentially, registration / titling of farms... along with the classic being done by ESRI, there are other new companies such as who are working on how to use imagery...

Weather and location: Companies such as Awhere is working with MicroSoft on systems that include providing farmers with information on best planting dates. This work uses intelligent cloud work, weather data and  geospatial analysis, that enables localised decision support for farmers.  

I believe aWhere is also using their geospatial tools to help organizations such as One Acre Fund to support the mechanics behind crop insurance support and potential triggers for insurance payouts when drought / flooding strikes.   

Logistics and assets / services: Hello Tractor is a company that is the uber for tractors..  so the software allows users in Nigeria to use their phones to locate and book a tractor at planting time..  Similarly, farmers are starting to use similar FIND applications to locate agro-dealers, or maybe compare prices across agro-dealers... They may not know they are using geospatial analytics, but they are!! 

Geospatial data

I agree there is quite a bit of interesting work around the use of geospatial data both for verifying production and potential shocks.  Thank you for your summary and I am sure there is increasing potential for partnerships between the various interested groups to take advantage of this data in order to inform policy decisions.

Topic locked