Hardware Considerations
The soil research survey reaffirmed our preference for 7-inch tablets over phones. The tablets allowed the team leader to teach others and show the day’s route to those gathered around. The tablets’ batteries made it through each day of research once we made sure to diligently use airplane mode and keep the screens turned off whenever possible.

Because we were literally out in the middle of open fields for much of the day, we also had to be much more conscious about not leaving the devices exposed to sunlight and the resulting high temperatures. Extreme heat can irreversibly damage the devices, but even more of a concern is the impact of heat on battery life.

The experience also reiterated the importance of the “jean pocket test.” In other words, look for devices that will fit in a back pocket of your pants. Anything bigger requires you to have a satchel, case, or bag with you at all times.

Order of Soil Research Questions
The soil research survey did not involve interviewing, as opposed to our socio-economic survey, which had. Instead, the soil data that was collected was observational, with many people making observations at one site that all needed to be recorded. With paper surveys, a team leader could jump around the paper filling out different sections as other team members called out information to her. But with the digital form, we needed to pay more attention to the order of the questions and the work flow of the team. Because we recognized this challenge early and planned for it, things went smoothly. Attention definitely needs to be paid when non-interview type surveys are converted from paper.

Using Data-enabled Devices/Challenges with Photos
We used data-enabled tablets (the type that take SIM cards) for this work rather than ones that only operate over wifi. This meant that we were able to send back data from our remote location without an additional Internet connection. The data speeds were very slow, and the cell towers were completely down for a full 24 hours at one point, but we were still able to send data from the devices to our servers.

We did, unfortunately, run into trouble trying to submit forms with photos attached. The data connection was too poor to transmit the photos, and the submissions timed out before they were sent. To solve this problem, we removed the photos from our survey, and took pictures using the camera app instead. We made sure the photos were geotagged, and we manually removed them from the devices instead.

GPS Accuracy
Generally Android devices and ODK Collect have trouble recording GPS points with less than 5 meter accuracy. Better quality readings are especially important for our soil research, so we recorded the same location three times within each survey. We have not yet determined whether that technique improved our results, but we will post here when we do.

Eric Couper presents at Columbia U.

Last week, Eric Couper, AfSIS’ ICT and Agriculture Coordinator, had the opportunity to discuss AfSIS’ mobile data collection work with colleagues and students affiliated with Columbia University’s Earth Institute. Chris Planicka wrote about the presentation for the New Media Task Force (NMTF) blog [link to post]. He summarizes the talk and encourages members of NMTF to consider the role Open Data Kit could play in their work.

Chris writes:

The results he shared at the colloquium highlight the possibilities, as well as the continuing challenges, of using mobile technology for data collection. We felt New Media Task Force members will find these results, and the hardware and software utilized by the team, interesting and useful!

The ppt from the presentation is available here.

Fig. 1. An example of deforestation of woodlands for charcoal production.

Sustainable management aimed at maintaining high amounts of standing biomass, reducing deforestation and re-afforesting marginal agricultural lands offer possibilities of reducing global warming and maintaining local and regional ecological systems. However, people and governments in developing countries do not have incentives to plan and enforce afforestation and forest protection measures vis-à-vis their demand for direct and indirect forest use. Recently, the United Nations Framework Convention on Climate Change (UNFCCC) has initiated discussions on mechanisms to Reduce Emissions from Deforestation and Degradation (REDD) in developing countries. The REDD concept proposed to provide financial incentives to help developing countries facilitate and enforce voluntary reduction of national deforestation and associated carbon emissions below a baseline. If properly executed, the emission reductions could simultaneously combat climate change, conserve biodiversity and protect other ecosystem goods and services.

To efficiently execute incentives and make sure that policies desired to tackle deforestation are implemented, knowledge about the current condition and extent of forest cover as well as the amount of carbon that was stored in those forests is necessary. It is also important to design monitoring mechanisms to assess trends and implement suitable rectification measures.

The main carbon sinks in tropical forest ecosystems are the living biomass of trees. The carbon stored in the above-ground living biomass of trees thus typically constitutes the largest pool and is the most directly impacted by deforestation and degradation. Estimating above-ground forest biomass is therefore one of the most critical steps in quantifying carbon stocks and fluxes from tropical forests.

Fig. 2. Areas with ‘biomass’ and ‘zero biomass’

Allometric equations are the most common approaches used to estimate aboveground biomass and carbon stock. Measurements to develop allometric equations could be achieved by direct and indirect methods. The most direct way to quantifying the carbon stored in aboveground living forest biomass is to harvest all trees in a known area, dry them and weigh the biomass. While this method is accurate for a particular location, it is prohibitively time-consuming, expensive, destructive and impractical for country or regional level analysis. Thus, developing tools that can help ‘scale up’ or extrapolate destructive harvest data points to larger scales based on proxies measured in the field or from remote sensing instruments is essential. The indirect approaches involve the estimation of difficult- to-measure parameters, such as stem volume and tree mass, from easy-to-measure tree parameters such as diameter at breast height (D) and tree height (H). However, since it can be time consuming to measure D and H for a large number of trees, other ‘indirect’ approaches utilize equations developed in similar environments and for similar species to estimate biomass.

The Africa Soil Information Service (AfSIS) project is well-placed to assess aboveground biomass of trees and estimate their carbon storage potential since it is acquiring a huge dataset for a large number and representative sites across sub-Saharan Africa (SSA). Among the numerous landscape attributes, the project acquired information on number of trees, D, and H, which can directly be used to develop site- and species –specific allometric equations, which in turn will be used to estimate aboveground biomass and carbon sequestration potential.

During the recent data analysis workshop in Arusha late November and early December 2011, project team members tried to make use of the available data to gain a general picture of aboveground biomass in Miombo woodlands and assess their carbon sequestration potential. We also attempted to assess the key drivers that determine biomass distribution, more specifically considering Miombo woodlands as an example.

Generally, biomass is a function of tree diameter at breast height, height of tree, and specific wood density. One of the most commonly used equations to estimate aboveground biomass thus takes the form:

          (1)

Fig. 3. The role of some landscape attributes in affecting biomass

where AGB = aboveground biomass (kg/tree); = model fitted parameter and varying across forest types; d  = specific wood density (g/cm³); D = trunk diameter (cm); and H = total tree height (m). The general form of allometric equation shown in [1] can be expressed in various forms mainly depending on the prevailing relationship between D and H.

In the Arusha workshop, the equation from Chave, J et al, 2005 (which is one example of the form shown in [1])  was employed to estimate AGB of 15 sentinel sites in four countries (Mozambique, Malawi, Zambia and Zimbabwe). Data collected using the T-square methods at subplot level (100 m²) were used for this analysis. Since there were many subplots with ‘zero biomass’ (Fig. 2), treating those in the analysis as ‘no data’ variables could overestimate AGB. As a result, we separated the analysis of biomass into two parts:

• Estimation of the presence of biomass for each subplot
• Estimation of biomass for places where it is present

The preliminary analysis results for part 1 presented here are thus based on the ‘dummy biomass’ variable created as the presence of biomass.

Though the equations used and models developed will require further testing and proper calibration, the results in Fig. 3 can be considered plausible. Further analysis is being conducted to compare different allometric equations and develop suitable modes that can help estimate Miombo woodland biomass as well as other forest ecosystems in SSA.

Fig. 4. Measured and predicted biomass for three examples sites (Gamb=Gambissa in Zimbabwe; Nkha= Nkhata Bay in Malawi; and Chic=Chica-b in Mozambique).

Figure 4 shows the correspondence between ‘measured’ and ‘estimated’ values, which generally shows very good performance of the model developed in this exercise. In all the cases (results plotted from three example sites in three countries) there is very good agreement between field-measured and predicted biomass for the Miombo woodlands in Southern Africa. Application of other alternate equations that have been developed for the region and more specifically for Miombo woodlands will further improve model performance.

Despite the fact that model-based biomass estimates show good agreement with measured values, the results presented in this study should be considered tentative, as this effort was only intended to demonstrate the possibility of using allometric equations to assess biomass and estimate carbon storage potential. Additional equations and models will be used to develop appropriate models.

Reference
Chave, J. and Andalo, C. and Brown, S. and Cairns, MA and Chambers, JQ and Eamus, D. and Folster, H. and Fromard, F. andHiguchi, N. and Kira, T. and others. 2005. Tree allometry and improved estimation of carbon stocks and balance in tropical forests. Oecologia. 145:1, pp. 87-99.

The hands-on session, which ran from Nov. 27 to Dec. 1, 2011, was hosted by AfSIS Project Leader Dr. Markus Walsh, and brought together scientists from AfSIS regional labs in Southern and Western Africa, CIAT-TSBF headquarters, the World Agroforestry Centre (ICRAF), the Earth Institute at Columbia University, and Ethiopia’s Agricultural Transformation Agency (ATA).

Participants first were given a chance to install and practice running the latest versions of R and Grass, widely used free and open source software for statistical computing and mapping, respectively. Dr. Walsh provided an overview of approaches to multilevel modeling and sampling designs, and Dr. Jiehua Chen, a statistician working for AfSIS, gave a talk on classic spatial analysis in R.

Drs. Walsh and Chen helped participants run new algorithms in R to generate soil carbon estimates for Western Kenya and analyze fertilizer response rates in Malawi. Participants also presented their own datasets and associated research questions, ranging from spectral characterization of fertile soil to biomass estimates for the Miombo woodlands in Eastern Africa to developing regional maps of earthworm prevalence.

Following these hands-on examples, participants formed subgroups to tackle some of their specific areas of interest. One group generated estimates of aboveground forest biomass using the field data collected under the Land Degradation Surveillance Framework (LDSF) used by AfSIS field teams. Because the main carbon sinks in tropical forest ecosystems are the living biomass of trees, such estimates are important in quantifying carbon stocks and fluxes from tropical forests. The carbon stored in the above-ground living biomass of trees typically constitutes the largest carbon pool and is the most directly impacted by deforestation and degradation — see this post for more details and early findings.

Another group linked spectral measurements to the Western Kenya data and generated predictions for additional soil properties, including pH and texture.

CIAT scientists also presented preliminary results from the socioeconomic and agronomic surveys undertaken in Western Kenya and worked with rest of the group on linking these datasets to the soil and remote sensing data for the same region.

See also:
Global Soil Map colleagues at the University of Sydney held a training program for West African soil scientists last month.

The six kernel density plots below show the distribution of key soil fertility constraints at AfSIS sentinel sites. The vertical lines in each plot show critical limits from left to right of low and high in the case of soil pH, very low and low for bases, P, K, and Zn, and low for Cu.

In the figure below, the laser diffraction particle size analysis of dry and wet soils is providing indices of soil erodibility. Soils in the left half of the graph are coarse textured soils and by definition do not give fine particles in wet or dry state. The soils in the bottom right quadrant are fine textured soils but are well aggregated in their dry state and so are not susceptible to wind erosion.

On the other hand, soils in the top right quadrant are fine soils that contain a high proportion of fine particles in their dry state (i.e. poorly aggregated) and so can be expected to be highly susceptible to both wind and water erosion.

On average about 50% of the samples received have been characterized with infrared spectroscopy and reference methods. The laboratory also provided primary soil processing services for four AfSIS Ethiopia sites, three western Kenya sites, and for diagnostic field trials from three sites.

AfSIS continues to send soil subsamples to Colorado State University and the University of Colorado, for DNA analysis of soil microbial biodiversity. Work is continuing in this area through further support from the Winslow Foundation, and being extended to additional soil faunal sampling at a selection of AfSIS sites with additional funding from the CGIAR Research Program on Water, Land and Ecosystems.

Training and capacity building

AfSIS laboratory procedures

During the past 12 months the Soil-Plant Spectral Diagnostic Laboratory received 151 visitors, provided exposure to the AfSIS laboratory methods to an additional 191 scientists and students, and gave more detailed instruction and supervision to 15 African PhD and MSc students.

A mid-infrared spectrometer was installed at the Kenya Agricultural Institute and a one-week training course provide to scientists and technicians from 14 to 18 November 2011. Further assistance is being given to KARI to establish an infrared spectral library based on their soil archives. A mid-infrared spectrometer was also purchased for Niger and will be installed din early 2012. Technical backstopping was also given to infrared spectral laboratories in Cote D’Ivoire, Cameroon, Malawi, Mali, Mozambique, and Tanzania.

A one-day training programme on AfSIS including a laboratory visit was given to 30 African scientists as part of a Kenya Refresher Course run by the Faculty of Geo-Information Science and Earth Observation (ITC) of the University of Twente, on 2 November 2011.

A seminar on AfSIS soil testing was given to 19 African participants of theInternational master Class for Africa on “The laboratory in the field: the practical use of new technologies in community-based extension”, organized by CABI, on 22 November 2011. A follow-up laboratory visit and testing of a field soil test kit was provided on 30 November.

A PhD scholarship was awarded to Stephen Ichami Muhati from the University of Wageningen on “Downscaling blanket fertilizer use recommendations for smallholder farms in African landscapes.” The study will test population-based sampling of farmers’ maize fields in conjunctions with spectroscopic analytical techniques for establishing soil and plant test critical limits and establishing prevalence rates of nutrient deficiencies.

Outreach

The Soil-Plant Lab also provided scientific and technical support to a number of projects in application of the AfSIS methods:

• The CGIAR Research Programme (CRP) on Water, Land and Ecosystems was approved in November 2011. The AfSIS methods have been proposed for global application across CGIAR sites under the Strategic Research Portfolio (SRP) on Information Systems and the support the SRP on rainfed systems. The methods will also contribute more broadly across CRPs to soil characterization at CGIAR long-term monitoring sites.

• AfSIS contributed an article to 2011 Annual Report for The International Union of Geological Sciences (IUGS)/ International Association of Geochemistry (IAGC) Task Group on Global Geochemical Baselines. We continue to seek opportunities for joint soil sampling and analysis through the IUGS geochemical mapping program in Africa and the report on AfSIS was designed to stimulate further interest.

• Further inputs were provided to a proposal for Integrated Monitoring System for Ecosystem Services in African Agricultural Landscapes being presented by Conservation International to the Bill & Melinda Gates Foundation. Inputs focused on providing a soils component to the proposed monitoring system.

• Sentinel sites are being established as part of a soil-plant health surveillance system for a Mars supported smallholder cocoa project in Cote d’Ivoire.

• LDSF baselines are being established and partners trained in FAO-MICCA projects in Kenya and Tanzania. FAO are developing proposals to extend the framework to projects in Ecuador and Vietnam.

• Kenya Forestry Research Institute are implementing a carbon inventory of Mount Kenya with the Kenya Wildlife Service. Soils from one sentinel site have been processed and three further sites sampled.

• A project to establish carbon and soil health baselines in Kenya rangelands was completed by Wajibu MS and ICRAF. Wajibu are engaged in proposals to extend the LDSF site network to conservation agriculture areas in Kenya and Tanzania, beyond their current pilot sites in Kenya.

• A BMZ-funded project coordinated by ILRI is underway to quantify land management effects on soil carbon in rangelands of Ethiopia and Burkina Faso. The project will establish sentinel sites and implement the LDSF in exclosures and long-term rangeland management experiments.

• The AfSIS soil carbon measurement methods were contributed to the UNEP-GEF Carbon Benefits Project: Modelling, Measurement and Monitoring (http://www.unep.org/climatechange/carbon-benefits/). The tools are expected to be applied in GEF and other sustainable land management projects.

Agriculture:
.

.

Project Description (from their Web   site):

…M-Farm, works as a transparency tool for farmers. [It]

1. Enable[s] farmers to inquire current market prices of different crops from different regions and/or specific markets
2. Aggregate[s] farmers needs/orders and connect them with farm input suppliers
3. Enable[s] farmers to sell collectively and connect them with a ready market

Biggest Take-away: Good farmer/market systems should do more than just provide price information. They should also connect farmers with interested buyers and help farmers sell collectively.

To Elaborate: M-Farm recognizes that knowing the price of maize at a particular market is often not enough. If it takes a farmer three days to get to that market, the prices might have changed or the buyer might be gone. Understanding that, M-Farm’s designers go a step farther: Registered users of M-Farm can request to buy or sell a particular crop, and M-Farm will provide information on corresponding sellers/buyers.

__________________________________________________________________

Project Description (from their Web site):

The ET LMIS is a mechanism through which collection, analysis and dissemination of information needed to help producers, middle men and traders are organized and systematized. This system provides near real time market information which is available on request via SMS text message system, email, WorldSpace radio systems and on the internet. The ET LMIS was created as part of the LINKS sub-project within the Global Livestock Collaborative Research Support Program (GL-CRSP) being implemented by Texas A&M University and funded by USAID.

Biggest Take-away: All cows are not created equal. In other words, systems dealing with livestock must address the complexities of animal varieties and grades/quality levels.

To Elaborate: LMIS is a comprehensive market information system. It captures information about livestock prices and makes it accessible through a variety of information channels (SMS, Web, radio). The idea itself is not so important, but what stands out is the designers’ knowledge of livestock trading. Knowing that the system cannot not simply capture <Cow sold for X Ethiopian birr>, they have built important information about the livestock variety and well-being into the system.

__________________________________________________________________

Project Description (from their Web site):

Farmer Voice Radio (FVR) brings a new concept of agricultural extension to serve the smallholder farmers of Africa.[ …] FVR provides full-featured, multi-media extension services based on radio, with active involvement of farmers and near-real-time program revision based on their feedback. Traditional extension models are based almost entirely on face-to-face farmer support. Systematic impact programming, with focus, depth, and repetition will allow farmers to make confident changes in their agricultural practices.

Biggest Take-away: FVR’s transition from long form agriculture programming to short, targeted, and repeated messages ( á la commercials) is more revolutionary than it sounds. It is important to break free of conventional thinking and formats for all of the information channels we use.

To Elaborate: Radio producers in the areas that FVR covers have had a very set opinion about how agriculture radio programming should be produced. They believed that programs should rarely/never be repeated and should be long form (30 minutes or an hour). After much debate, FVR is now developing “AgTips,” shorter segments that are regularly repeated. By keeping the medium (radio), but changing the format (from 30 minutes down to 5 minutes), FVR believes its messages will have a greater impact.

__________________________________________________________________

Other Sectors:

Project Description (from their Web site):

ColaLife is working in developing countries to bring Coca-Cola, its bottlers and others together to open up Coca-Cola’s distribution channels to carry ‘social products’ such as oral rehydration salts and zinc supplements to save children’s lives. [...]

ColaLife started as an online ‘movement’ in April 2008. We have more than 10,000 online supporters and these have given us the power to engage Coca-Cola, UNICEF and other key stakeholders. We are now focussing on getting a trial of the ColaLife concept underway in Zambia

Biggest Take-away: The best way to understand how to address the “last mile” –the last part in the process of delivering connectivity from communications provider to customer–is to consider what/who is already getting there.

To Elaborate: ColaLife builds on that idea that Coca Cola is available nearly everywhere on the planet that people live. It recognizes the success and power of another group and seeks to harness that as a resource. ColaLife’s idea challenges everyone serving rural populations to take lessons from and partner with organizations (for- or non-profit) that are already covering the last mile.

__________________________________________________________________

Project Description (from their Web site):

Young Africa Live consists of a combination of regularly updated, dynamic stories and live chats and a series of permanent content pieces. …The permanent content covers the essential facts around HIV and AIDS; as well as helpline numbers and contact details for referral organisations who can support YAL-users in times of need. Under the dynamic content section, daily news & celebrity stories are posted and users can ‘like’ the story or leave comments. Guest bloggers write about their experiences with love, sex, dating, cultural dilemma’s, gender stereotypes and other relevant issues. In addition, live chats with doctors and relationship experts help create a true dialogue between the portal and its users.

Biggest Take-away: The most direct, unadorned, way of conveying information is not always the best way. Content packaging (making it cool, exciting, participatory, and providing a context) is just as important as the content itself.

To Elaborate: Traditional behavior change messaging involves blasting a recommendation/fact/warning onto billboards, backpacks, and bumpers. Even the more nuanced messages (wrapped into soap operas and children’s stories) do not operate on the same level as Young Africa Live. It delivers content that first and foremost is interesting, and as a result has developed a community that genuinely engages (anonymously when desired) with important issues.

__________________________________________________________________

Project Description (from their Web site):

Ummeli is a mobile network that helps communities create their own employment opportunities. [...] Ummeli provides tools to a variety of key players in the jobs market:

• People who live in developing countries with little or no formal employment.
• Companies who seek temporary or long-term employees.
• Communities who want to build projects to improve their environment.
• NGOs looking to find volunteers.

Biggest Take-away: Existing, popular social networks are incredibly useful for launching new services.

To Elaborate: Ummeli is brought to us by the same group, the Praekelt Foundation, as Young Africa Live (YAL). The Ummeli team was able to feature and promote the new tool on YAL, and that helped expose Ummeli to a very large audience very quickly. Ummeli’s spot-on identification of a real need and well-designed service cannot be understated, but it is really rare in ICT4D work to see one service successfully promote another at this scale.

Categorizing Existing Projects/Tools
To start the conference off, SANGONET and Dalberg provided a framework for considering the role that different projects and tools play. Given the focus of the conference, they highlighted the education, agriculture, health, and governance sectors. They divided work in those sectors by their purpose (data collection, information/promotion, training & diagnostic support, service delivery, reminder/behavior change support, and financial services). The majority of AfSIS’ ICT4Ag work to date would fall in the Agriculture/Data Collection square, with some of our navigation work falling into Service Delivery.

Their introductory document (download as pdf) includes this framework (on page 5) and a lot of other interesting analysis.

Common Themes/Ideas
Charging for Services. The majority of individuals participating in the agriculture sector discussions agreed that agriculture services should not be given away. The participants want to reach scale with their services. They noted that this is particularly possible in the agriculture sector because it is reasonable to request some payment for services that either increase a farmer’s income or save her money.

Behavior Change Communications (BCC). Whatever the medium of information delivery, many ICT4RD projects hope to influence and/or change the behavior of rural Africans. While this was not always explicitly discussed, it was an undercurrent of almost every conversation. Within BCC, we noticed that initiatives operate somewhere along a spectrum. On one end, projects directly advocate specific changes (Farmer, you should do X!), and on the other, projects advocate for behavior change with more nuance via information (Farmer, did you know X? Perhaps that might influence your decisions?…) Regardless of the approach, nearly everyone was interested in the best approaches for communicating with farmers and influencing their behaviors.

AfSIS Contributions
We had the opportunity give two presentations at the conference. Both presentations seemed to be well received. If you were present and have feedback, please let us know in the comments below or email Eric at eric.couper (at) gmail (dot) com

Both sessions dealt with streamlining operations (saving time and money) rather than changing paradigms.  In the first presentation we discussed our use of Open Data Kit. The presentation was a condensed version of our writings here with an additional twenty minutes set aside for a hands-on demonstration on how to use ODK. The demonstration helped the participants realize that ODK is not as complex as it initially seems. At least one organization left the presentation thinking seriously about adopting ODK for their work.

• Presentation (.pdf)
• Demonstration Form customized for conference (xml, .xls)

The second presentation discussed our use of Locus and Open Street Maps. The content was very similar to what we discussed in this post. We walked through the tools that AfSIS has been using, and together we brainstormed ways to improve our organizations’ transportation systems. For example, one person recounted a famous pink building that her organization has used for years as a landmark. The building was recently painted another color, and she spent hours trying to find the route! In the future, that organization will not abandon landmarks completely, but they are considering more technical approaches as a supplement.

—-
Special thanks to Bart Sullivan of Farm Radio International for recommending the conference and helping us get presentation time on short notice.

Imagine that your organization is working in a new area for the first time. If you are working in the United States, you can most likely open up any number of online map services to learn about road density, quality of roads, other important businesses and landmarks in the area, etc. If you are working in rural Tanzania, this is much less likely.

Without detailed transportation maps, your organization might lose time and money, getting lost, encountering impasses, getting stuck, etc. To make it worse, the lessons learned by one person in one vehicle are not easily transferred to others. Luckily there are cheap and easy solutions to manage this challenge.  For our work, we have been using the same Android devices used to collect data and the Locus Free application we were already using to navigate to particular points in the field.

AfSIS procedure

• Using our Android devices and the Locus Free app we recorded the route our vehicle took each day.
• Open Locus > Locus button (in top left corner) > Functions > Track record > Start/stop/save
• Starting one the second day, we could view the tracks we recorded and use them for navigation
• After a few weeks, we downloaded the .kml track files (click to download) from the devices and uploaded them to Open Street Maps

Advanced procedures

• During our first pilot, we were only using one vehicle/driver, so sharing tracks between devices was not necessary. For projects using more than one vehicle, it might help to share track files between devices.
• We also simply recorded our track. If a track/road ends abruptly, we assumed that the road was impassable and avoided it. In the future we might label obstacles (no road continuing, tree across road, too muddy to continue, etc.). Doing so would provide helpful context. As an example, a road that is too muddy to drive on in the rainy season might be great to use in the dry season.

Adding Open Street Maps
As described on their website,

“OpenStreetMap is a free editable map of the whole world. It is made by people like you. OpenStreetMap allows you to view, edit and use geographical data in a collaborative way from anywhere on Earth.”

Used well, OSM can be incredibly useful and add even more value to the track files that you record.

Comparison Map. Before AfSIS starting working in the Kisongo, Musa, and Mateves wards, just west of Arusha, Tanzania, these two maps looked very similar. AfSIS was able to contribute the dotted brown lines that represent previously unrecorded dirt roads.

Why bother?
Admittedly, your organization does not have to contribute its tracks to Open Street Maps to take advantage of the tracks it records, but doing so can be helpful for you, other organizations, and, most importantly, the community you are serving.

Your organization can save and use your road data without uploading it to Open Street Map, but that assumes the data will be managed well and passed on from year to year and from person to person. The easiest way to make sure the information is still available in the future is to upload it to OSM.

Other organizations face the same challenges that your organization does, and there is no reason why we should all re-track the same roads, get stuck in the same mud pits, and get lost at the same intersections. Your organization’s minor efforts can save another from tremendous headaches.

An unmapped community is almost always an underrepresented community. Road maps (even of dirt roads) represent an opportunity for increased access to public and private goods/services. Road maps often influence government policies and private business strategies in the same way that the census data and government registration can. People always count, but they can more easily access resources when they are counted. Road mapping is an important part of that process.

Links to Open Street Map resources
Beginner’s Guide
Viewing OSM
Available map features, example of dirt road types

Purpose
There are a variety of reasons an organization might develop a form beyond the one being used for research. Reasons include:

Training on ODK. When someone uses ODK Collect for the first time, it is probably a bad idea to give her your long and complex form to flip through. Instead a training form or forms can slowly introduce trainees to different features of ODK (skip logic, picture, GPS, etc.) without overwhelming them.

Controlling Quality. Given all the things that could go wrong with a device each day, it is wise to check that the devices are functioning properly. A simple test form submitted each morning before enumerators begin their trainings will go a long way.

Getting Feedback. Ideally everyone involved in your data collection is comfortable discussing the challenges they are facing and the issues that are frustrating them, but this is not always the case. Meta-forms that assess employee satisfaction can provide an anonymous outlet for enumerators to share and give suggestions.

Piloting Hardware. Very often organizations test out the benefits of ICT using a pilot, testing out what works and how well. Meta-forms can be used to track these experiences. For example, a meta-form might capture the battery life remaining in each device at the end of the day, or gauge enumerators’ device preferences.

Our Downloadable Forms
(N.B. Right click on linked files, then select “Save file as…” to save)
Battery Life. (xml, xls) To measure the battery life of each device at the end of the day. We were tracking battery life to vet new devices. Used daily.
Enumerator Training Experience. (xml, xls) To gauge the enumerators’ experiences with our training. Used at the end of training.
Enumerator Weekly. (xml, xls) To gauge enumerator sentiment, capture anonymous discontent, etc. Used weekly.
Training Practice Survey. (xml, xls) Survey used during training to familiarize our enumerators with ODK. Used during training.

Fast Download: All files saved together in zipped folder.