Poverty Intelligence Network « Nathalie Collins

Poverty Intelligence Network

During the summer of 2010, I worked alongside the members of the Kurian community of western Kenya to develop a mobile phone based intelligence network. Its purpose was to use information collected by Nuru International’s field staff in their daily work to inform community development initiatives in the area. By asking a few simple questions during routine home visits or at local water sources, we were able to paint a living picture of the community’s progress out of poverty.

Nuru is a holistic community development organization committed to equipping the poor in remote, rural areas to end extreme poverty in their communities within 5 years. The organization is based on a service leadership model which trains and mentors local leaders in the community to create & sustain effective interventions in Agriculture, Water & Sanitation, Healthcare, Education, and Community Economic Development. As part of this effort, Nuru is investing in ongoing data collection & measurement initiatives to better understand the progress the community is making.

The Project
We started a project dubbed the Poverty Intelligence Network (or PIN) with the intention of using mobile phones in the field to collect ongoing information about its initiatives as well as observations about the overall community. The goal is to get a living pulse on the community which will enable rapid decision making and response. The system is based on low cost phones which are distributed to Kenyan staff members in the field. These are basic internet-enabled handsets with an onboard camera capable of taking both still pictures and short video. The phones are distributed pre-configured with Opera Mini 4 and are programmed to access a password-protected Google site containing a set of Google Forms based surveys.

The phones are being used as communication tools across Nuru’s program areas with initial data collection applications in the areas of Agriculture, Water & Sanitation, and Healthcare. Over the last couple of months, the team has focused on three main initiatives:

Conducting a comprehensive census of the project area in order to create the first accurate map of the region
Correlating census data with locations of and qualitative interviews at existing water sources in order to plan for appropriate water & sanitation interventions
Mapping the locations of CHWs in order to determine their distribution across the households in the community

We have used the phones in conjunction with separate GPS units in order to document the location information used in these applications. The data has then been analyzed either in Excel or imported into Google Earth to create visual renditions of the project site which have then been used in operational decision making.

Intentions
Coupling low cost handsets (Nokia 1680c priced at <$40 USD) with existing data collection mechanisms such as Google Forms gave Nuru a plug-and-play method for documenting community life. Not that there weren’t early bugs in assuring the appropriate version of Opera Mini would consistently access and submit forms, but there’s a certain desire in this space for a customizable solution which doesn’t require an army or even a handful of programmers to create. Even though Google Forms requires connectivity to complete surveys, early spot checks showed that Safaricom network service was widely available in the project site so coverage should be good. Calculations showed that relying on GPRS transmission (~$0.10 USD per MB at hundreds of form submissions per MB) over SMS ($0.03 USD per message) was much more cost effective. Google also offered an online form designer which was user-friendly enough to be used by the broader staff. This all looked very promising, so we set out to put it into practice.

Early Learnings on Phone Use
Once the system was set up and running, we equipped the staff with phones, trained them on their use, and then proceeded to observe how they were being
applied in the field. Some very interesting, and broadly applicable, learnings emerged:

Mobile phones are surprisingly stealthy interviewing tools. Given our own experiences with the social dynamics surrounding mobile phones, we expected their use to be a potential distraction during the interview process. We came to find that documenting conversations using a phone rather than paper & pen actually put the interviewees at ease. Feedback from our Water & Sanitation field staff was that carrying a notebook would often cause those being interviewed at local water sources to become uncomfortable while they didn’t initially suspect the mobile phone as a tool for note taking. This led to much more natural conversations between the staff and the individuals being interviewed.
Network coverage is not as ubiquitous as we thought. Though service in our very rural project area is actually quite good, we found there to be enough localized dead spots to prevent the staff from effectively completing the online Google Forms. In many cases, they would resort to taking paper notes which would then be later entered into the phone interface. This became the top issue cited in feedback on use of the system. More on this in the next steps noted below.
A method for data validation is absolutely key. We found that the staff expressed surprisingly little frustration with entering data using a basic keypad. However, the results of their submissions required some scrubbing – blank fields, typos, or duplicate submissions as they continued to learn the best way to complete the forms. Any system for data collection needs to account for this – conversations with some of you in this group have yielded similar observations.
The phones quickly grew legs of their own. Perhaps not too surprising but at least somewhat unexpected, we found the staff discovering a variety of applications for their new phones. In addition to using them as survey tools, many discovered the value of the calculator, scheduling notes in the calendar, and the ability to access online information. This may sound basic, but was fascinating to see given that this group initially spent hours training on basic keystrokes before being sent out into the field. In addition to surveying, the staff is getting their first glimpse of a worldwide ecosystem of information. This can be viewed as unrelated or complementary, but is something to consider for any project which equips rural populations with mobile technologies.

Initial Results + Next Steps
The initial results of the three initiatives mentioned above have been quite promising. The census is well underway and through the use of a basic GPS + an internet-enabled mobile phone we have been able to collect the location information of over 700 households. We’ve paired this information with feedback from qualitative interviews at local water sources to create a visual representation providing initial clues on how choices on where to fetch drinking water are being made. The location of CHWs in the community mapped alongside households is also giving us a definitive way to distribute them across the community so they can be most effective in their visits.

Given the early learnings described above, we are also working to make adjustments in this system to best suit the needs of data collection in rural areas. Though inspired by the plug-and-play nature of Google Forms and convinced of the cost-effectiveness of GPRS over SMS, we need to adjust for the observed spottiness of network connection. We are now experimenting with the use of OpenXData on these phones in order to provide a reliable means for offline data collection. We’ve realized there is a huge need for an easily configurable data collection solution that runs on low cost phones and though Google Forms could work well in more urban areas with consistent network coverage, it’s not as well suited to our project site. I’m eager to see more options geared toward non-technical NGO users which provide equivalent ease of setup.

In addition to the very practical matter of solidifying the basic technical components in the field, Nuru is looking ahead to applications of the Poverty Intelligence Network far beyond simple data collection. The ultimate vision of the PIN is to paint a living picture of the community which will enable rapid decision making and response. Can we use observations of increased temperature and diarrhea rates across a given area to quickly diagnose and respond to disease epidemics? Can we correlate changes in agricultural yields with school attendance and child malnutrition rates? Learning how decisions are made regarding where water is fetched or what type of seed is planted, can we provide training to the community which will help them to make good choices? And in addition to all of this, can the phones themselves become vehicles for distributing this knowledge? These questions and many more are front of mind for our team as we continue to prototype data-driven solutions in this space.

Additional Reading
For additional background on the PIN project, please take a look at the following set of blog posts:

Nuru’s Poverty Intelligence Network