Analytics Revolutionizing International Development, Disaster Recovery

by   |   December 10, 2014 5:30 am   |   0 Comments

Mother of eight Risper O.*, from Sindo, a small town in southwestern Kenya, is beaming with pride. In less than a year she has already managed to repay 15 percent of the US$200 microcredit with which she purchased her three beehives. Once she repays the entire loan, she expects to pocket around US$34 per harvest, with which she will be able to purchase a couple of pairs of new shoes and textbooks for at least one child, and pay for basic medical checkups.

Risper is one of the 2,735 East African smallholders who have joined the global honey market thanks to Honey Care, an inclusive business that provides impoverished communities with a source of income through sustainable beekeeping.

In December 2013, Risper received a visit from one of Honey Care’s loan officers, who logged her basic information into a smartphone using TaroWorks, a software application used on a mobile device such as a phone or a tablet, that was developed by the Grameen Foundation and was launched in January last year.

As well as recording Risper’s basic information, the loan officer asked her 10 questions about her standard of living, including what materials her house was made of and whether she had a latrine. This survey is Grameen’s Progress out of Poverty Index (PPI) and was designed for NGOs and development agencies to monitor changes in their beneficiaries’ quality of life during the course of a project, to ascertain whether the desired impact has been achieved.

Through the TaroWorks platform, Risper’s details were uploaded onto the Salesforce platform, a popular customer relationship management (CRM) platform that is currently used by businesses throughout the world.

Another member of Honey Care’s staff accessed Risper’s details through the Salesforce platform and paid her a second visit to determine whether her farm had the necessary elements for beehives to thrive, such as access to water and an abundance of flowers for bees to pollinate.

The information collected during the site assessment, including the size of Risper’s farm, a GPS tag recording its exact location, the presence of indigenous trees, and the type of crops that she grows, was logged into TaroWorks. If connectivity was poor, Honey Care’s staff could save the data in their phones and uploaded it once they reached an area where Internet access was available.

Back in the office, once Honey Care established that Risper’s farm was suitable for beekeeping, the information recorded by the loan officer during the first visit was used to create a profile for her on Kiva, a global online lending platform. The profile included a photograph and a short biography describing what Risper intended to use the money for. By January this year, Risper had received a total of US$200 from seven lenders, which she used to purchase her three hives.

“All of this used to be a big headache because we had to carry lists of farmers that the managers had to check and then report back to the salespeople, which could be very confusing,” said Honey Care director Andrew Loebus. “We were shuffling paper around and making lots of phone calls, which was very time consuming.”

Related Stories

U. of Tenn. Students Use Analytics to Assist Nonprofits.
Read the story »

What Businesses Can Learn from Public Safety Analytics.
Read the story »

When Everything in a Natural Disaster Revolves Around a Map and GIS Data.
Read the story »

KPI of the Planet: Integrating Sustainability into Corporate Decision-Making.
Read the story »

Beyond the obvious advantage of saving time and resources, TaroWorks has also allowed Honey Care to process the data collected during the site assessment to establish correlations between productivity levels and variables such as the proximity of the beehives to Risper’s homestead and the type of trees on her farm.

Salesforce uses TaroWorks to transform the data into charts and graphs that make it possible for Honey Care’s staff to spot trends, patterns, and correlations that will guide future actions.

“We’re going to make decisions about which regions are more or less productive and per-farm disbursement, so if farms that have a very high number of hives are less productive, it’s a good indication to be careful with how many hives we put in a piece of land,” said Loebus. “We’re also going to look at foliage and flowers to see if there are correlations between productive hives and foliage crops on the farms so that we can recommend planting certain things to farmers.”

Customer Segmentation Applied to Development

TaroWorks was created in response to the Grameen Foundation’s need for better information about the impact of its projects. “We were running our own interventions in the field and we realized that we needed real time access to what was working, what wasn’t, and where our field staff was,” said TaroWorks Executive Director Emily Tucker. “If we’re going to be successful, we need to see what’s working in two weeks or in one-month cycles so that we can iterate quickly, roll out, and deploy the things that work and discard those that don’t.”

The Grameen Foundation engaged in a series of talks with more than 225 NGOs, found that they all had very similar needs, and decided to create the TaroWorks tool and make it available to other organizations.

Tucker drew from her past experience in the retail sector, where customer segmentation and understanding repeat usage behavior have become standard practice, and applied these concepts to the development field.

Trocaire, the Overseas Development Agency of the Irish Catholic Church, is one of the 43 inclusive businesses and NGOs across the world that are currently using TaroWorks. In 2013, it used TaroWorks as part of a pilot project with four partner organizations to measure the impact of its Sustainable Livelihoods project in Malawi.

In October this year, it drew up an internal report that summarizes the lessons learned from this experience. Overall, the results were positive, although the document also highlights a number of areas where there is room for improvement for both Trocaire and the creators of TaroWorks.

Over the past five years, Trocaire, which works in more than 20 countries, has become increasingly driven by results-based management, which has generated a far heavier data burden for the organization. Trocaire had already used Salesforce and decided to give TaroWorks a try in the hope that information technology and data analytics could help to ease this burden and improve efficiency.

“The expectation is that we will have more consistent indicators across programs and that we can look at them and see how a particular country is doing and what lessons can be learned,” said Maria Collison, Trocaire’s monitoring and evaluation officer. “It’s about standardizing and having data available and actually used as opposed to having big evaluation events with data sitting in Excel spreadsheets and never being used after that.”

The report concluded that, thanks to the use of TaroWorks, the efficiency with which data was collected and processed definitely improved. On the downside, it found that the data generated by the PPI could not be used because the index was based on outdated standards of wellbeing that had been established in 2005.

Added to this, the fact that information technology and mobile devices are still not readily available in Malawi made it hard for partner organizations to feel at ease with the tool. Also, a focus on the technology itself detracted from soft skills, such as interview techniques, which had a detrimental impact on the quality of the data collected.

“The technology works in a vacuum, but it doesn’t solve all of your problems because you don’t work in a vacuum,” said Collison. “Maybe there was a lack of capacity from Trocaire compounded by the fact that trying to conduct a survey when you’re not extremely good at this while trying to operate a phone can be distracting.”

‘It’s about standardizing and having data available and actually used as opposed to … data sitting in Excel spreadsheets and never being used after that.’

Collison added that the fact that TaroWorks, like any other software, is constantly updated meant that partner organizations had to be re-trained on how to use its latest version halfway through the pilot project. Nevertheless, Collison believes that some of these issues are merely teething problems and said that Trocaire is likely to conduct other pilot projects in Africa.

To address the issue of soft skills, Tucker said that TaroWorks now includes reminders for interviewers such as the importance of ensuring that the physical space where the interview is conducted is appropriate, explaining clearly what the data will be used for, and thanking the interviewee at the end of the survey.
The next version of TaroWorks will include new dashboards, mapping functions that will allow NGOs to track field agents over a particular period of time to ensure they are servicing their designated areas adequately, and a scheduling function that will enable staff to assign particular tasks to each field agent.

Data Analytics for Disaster Response

Typhoon Haiyan, one of the strongest cyclones ever recorded, hit the Philippines in November 2013, killing at least 6,300 people and affecting 11 million, many of whom were left homeless.

For the International Organization for Migration (IOM), a global inter-governmental organization that provides disaster relief for people displaced by natural disasters and armed conflicts, processing data in a quick and efficient manner to establish what relief was needed and where was of paramount importance.

The IOM used a platform developed in 2004 to assist the victims of the Iraq war, called the Displacement Tracking Matrix (DTM), that gathers data such as access to water and the health and sanitation conditions in the evacuation camps. It used software donated by SAS to process the data and analyze patterns and trends such as the number of latrines per capita in an evacuation camp.

Phyo Wai Kyaw, the IOM’s information manager in the Philippines, explains that SAS analytics software made it possible to carry out a sentiment analysis and classify the comments made by evacuees as positive or negative on a relevant scale, in order to obtain valuable qualitative information.

“You can pull out words that get mentioned quite a few times,” said Wai Kyaw. “That’s data that we can merge or match up. Before, that was impossible. SAS also has quite a few charts, the traditional ones and the new ones, and you can combine them together. Among them are geo bubble maps that can be mapped directly on Open Street Map. There’s a lot of functionality in the system and I think we’ve only scratched the surface.”

The IOM also used SAS to carry out a text mining analysis of more than 10,000 tweets that indicated total structural devastation in Guiuan and revealed that the supplies that local hospital needed most to meet increased healthcare demands were essential medicines such as antibiotics and fuel for generators.

The text analysis revealed the most common health complaints to be upper respiratory and cold symptoms. But the most alarming findings were the high concentrations of diarrhea, fever, and skin disease among older people in evacuation centers in Leyte. This data that was shared with local health authorities to address these health needs.

“The software makes it a lot easier to discover those correlations and then do the analysis. You can filter to do the analysis with Excel, but it would have taken a lot longer,” said Lorelle Yuen, the IOM’s DTM rapid response officer.

On the downside, Yuen said that the software is very heavy, which has made it difficult to use in countries such as South Sudan, where connectivity is poor.

But overall, Yuen said that the IOM regards data analytics as the way forward. The IOM, which already uses tablets to collect data, is currently developing a mobile application and is exploring the possibility of using predictive analytics.

Another important lesson from Typhoon Haiyan was the above-average number of local volunteers who came forward to take part in the relief effort. Patrick Meier, an expert on humanitarian technology who has worked with the UN and other organizations, is currently trying to tap into this resource by creating an interactive platform called “Zoomanitarians,” where before-and-after satellite images of some of the areas most affected by Typhoon Haiyan are being uploaded so that volunteers from any part of the world can identify damage to infrastructure, debris, and blocked roads. If the experiment works, the platform will be ready to use in real time, in the event of a natural disaster, by mid-2015, Meier said.

The Power of Big Data

In response to United Nations Secretary General Ban Ki-Moon’s call for a data revolution in 2009, the UN created its own data innovation lab, Global Pulse. Last month, Global Pulse’s newly established Independent Expert Advisory Group (IEAG), which brings together a wide range of experts in information technology and development, produced a report titled, “A World That Counts. Mobilizing the Data Revolution for Sustainable Development.” In this document, the IEAG explains how the use of data that was driven by the adoption of the Millennium Development Goals (MDGs) in 2000 will intensify as the world adopts the Sustainable Development Goals (SDGs) in 2015.

The problems identified by the report include inequalities in terms of access to data and the ability to use it, as well as the fact that data often is released too late or not at all, and frequently is not well documented or harmonized or not available at the level needed for decision making.

In an attempt to overcome some of these problems, UN Global Pulse has partnered with SAS as well as social media analytics companies Crimson Hexagon and DataSift to process a wide range of nontraditional data sources, including mail flows, cell phone data, social media traffic, and blogs.

“The next wave is not just providing data to the masses, but delivering statistical and analytical insights to the world,” said I-sah Hsieh, global manager for SAS’ international development division. “We need to present the data in a way that allows people to drill down, slice and dice data, and present it in the way most relevant to them.”

Using SAS analytics software, Global Pulse monitored more than 200,000 online documents in Indonesia to extract relevant information about food prices, fuel prices, and unemployment. SAS software was used to detect the location, price, and availability of specific government programs, analyze sentiment and mood of social media content, and condense the results into a series of graphs, statistics, and charts.

Social media data was correlated with macroeconomic and unemployment metrics, and SAS identified leading and lagging indicators of trends such as inflation and unemployment. The data was integrated with offline attributes such as official statistics and other online metrics into policymaker-centric criteria, thus giving policymakers information they can act on.

‘The next wave is not just providing data to the masses, but delivering statistical and analytical insights to the world.’

Global Pulse is currently implementing a similar research project in Egypt, where Twitter footprints are being analyzed to establish whether there is a correlation between anxiety expressed by social media users regarding the possibility of losing their jobs and official unemployment rates.

“If we find some level of correlation between Twitter signals and the ground truth, our next step is to develop a platform that allows the United Nations Development Program (UNDP) in Egypt to monitor in real time that particular index on social media,” said Meier, who is leading the UN’s research project in Egypt. “These are all geo-located tweets, so if an intervention is carried out in a particular area and three or six months later the index bumps in that area and we can relate that back to the intervention, we can develop a decision support tool based on the research tool.”

Over the next two years, Global Pulse will focus on data related to gender, climate change, and resilience, as well as frameworks to protect privacy when working with big data. New types of data, such as digitized radio content, postal flows, and high-resolution satellite imagery, also will be included in future projects.

“Traditional information and statistics are like taking a high-resolution photograph every once in a while, whereas this new era of digital data is generating something that’s always pulsing,” said Anoush Rima, Global Pulse’s strategic communications and partnerships officer. “It’s like taking grainy video footage of something over time and, in between, taking that high-resolution photograph. Analytics allows us to filter and make sense of the data in order to gain new insights into people’s changing priorities at a scale and get insights that maybe weren’t available before.”

And it’s not only large organizations such as UN agencies and international NGOs that are using these non-traditional sources of information. Small, community based organizations are increasingly tapping into these sources as well.

The Inter-Ethnic Association for the Development of the Peruvian Jungle (AIDESEP), for example, is currently using drones to monitor potential threats to indigenous communities, such as illegal logging and mining, as well as land invasions. Drones, initially developed by the military intelligence sector, have become a widely used data-gathering tool for large corporations such as Amazon and United Parcel Service (UPS).

“This is another data-collection tool that can empower local communities so they can gain agency and potentially change the balance of power by providing evidence to counter government narratives,” said Meier. “Drones are becoming cheaper every month and they have a democratizing potential.”

Rob Nachtrieb, a professor at MIT’s Sloane School of Management, explains why extracting crucial information from a vast ocean of big data can be crucial for helping development agencies and NGOs to understand the root causes of poverty and plan more effective interventions.

“Poverty is so hard to get rid of because it’s a structurally stable system,” he said. “There are several reinforcing feedback loops and balancing feedback loops that make the system, so if you just change one thing, the system has a tendency to push you back to where you started. It’s like a nest of cause-and-effect relations, but if we can understand what those cause-and-effect relationships are, we can find a leverage point that might be particularly effective and our attempts to alleviate poverty might have some impact.”

Big data, says Nachtrieb, can be “a toolbox” that can help to solve these problems using the scientific method,  meaning an understanding of cause and effect relationships between different variables.

“You can collect data and statistics and, through the application of systems dynamics, you can hypothesize these relationships and see how the system would behave if you made an intervention. Big data is helpful to look for correlations and confirm that the connections actually exist,” Nachtrieb said. However, he warns that “big data is a tool but there isn’t a single silver bullet to combat poverty effectively.”

What NGOs Can Teach the Business World

The use of data analytics in the nonprofit world is a relatively new trend that began to gain momentum only two or three years ago. “The sector as a whole is lagging years behind what’s happening in the private sector. Just now it’s beginning to wake up and realize the power of using data to drive the work we do,” says TaroWorks’ Tucker.

The data revolution that has swept across the global development sector in recent years, Global Pulse’s Rima said, is likely to bring about “a cultural shift” in nonprofit organizations, as they increasingly switch to hiring professionals with backgrounds in mathematics and computer science who can maximize the potential of analytics technology.

But although nonprofits have borrowed the use of data analytics from the business sector, development experts point out that the business sector can learn important lessons from the unique way in which nonprofit organizations have adapted analytics tools to the context of development and poverty alleviation.

“SAS develops data analytics tools, but does it know every single way in which they’re used?” said Nachtrieb. “I think there’s an opportunity for learning to occur in both directions.”

One of the key lessons to learn, Meier said, is the importance of developing technology that is highly resilient and can be used in challenging environments.

“In the humanitarian technology space, we are experimenting with prototypes of new technologies under very extreme, difficult circumstances,” said Meier. “If a piece of technology works in the middle of a disaster, it will work anywhere else.”

* For confidentiality reasons, Honey Care prefers not to disclose Risper O’s full name.

Louisa Reynolds is a British journalist who has reported from Latin America for the past eight years. In the 2014-2015, Louisa was awarded the International Women’s Media Foundation’s Elizabeth Neuffer Journalism Fellowship and as part of this program she is spending six months at MIT as a research scholar and is writing for the Boston Globe and The New York Times.

Initially spending two years working as an editor for Central America Report, she later joined El Periódico daily newspaper, where she spent her early days covering business and finance and later switched to politics. Freelancing since 2011, Reynolds has worked as a local editor for the Investigative Reporting Initiative in the Americas, a project supported by the International Center for Journalists (ICFJ).

Her articles have been published in a range of local and international publications including The Boston Globe, The Economist Intelligence Unit, Inter-Press Service, Latin America Press, Latin America Database, Plaza Pública, Estrategia & Negocios, El Periódico, Siglo 21, Proceso, Latam Investor, Insight Guyana, International Justice Tribune, Jane’s Foreign Report and Global Radio News Live.

In 2013, Louisa won the Gold Standard Award for Excellence in Economics and Business Journalism for an article on inclusive and sustainable corporate models published by Estrategia & Negocios magazine.

Louisa has a bachelor’s degree in Modern Languages from University College London and a master’s degree in Latin American Studies from the Institute of Latin American Studies, as well as a post-graduate diploma in Newspaper Journalism from Lambeth College.

Subscribe to Data Informed for the latest information and news on big data and analytics for the enterprise.

Tags: , , , , , , ,

Post a Comment

Your email is never published nor shared. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>