Dr. Xiao Hu of the University of California's Department of Neurosurgery in Los Angeles analyzes brain waves to predict the rise of deadly brain pressure. Predictive analytics can also help the global development community target its resources, argue Susan Hayashi, Manu Singh and Radha Gholkar in a guest commentary for Devex. Photo by: ibmphoto24 / CC BY-NC-ND

Innovation is a word we hear all the time. Whether we work in the nonprofit or for-profit sector, in business or in international development, in government or at the grassroots level, everyone talks about it. Here is an opportunity to do it!

At JBS International, we recognize that the challenge is to understand the difference between fostering innovation and simply throwing money at a problem or a new idea — and hoping it works.

As the field of international development looks for new ways to innovate, program practitioners — planners, implementers and evaluators — as well as policy-makers and donors need innovative tools of their own to leverage the information they have. Traditional program research and evaluation efforts are sometimes costly, too far removed, and difficult to translate into effective planning. Wouldn’t it be great if we had cutting-edge tools to enhance real-time, data-driven decision-making and address the fast pace of humanitarian and development work, where decisions about human lives and precious resources are at stake?

That’s where predictive analytics comes in and can more directly target the efforts made to increase women and girls’ rights, empowerment and opportunities as an integral part of the global development landscape.

What is predictive analytics?

Predictive analytics is a state-of-the-art group of data analysis techniques that are designed to look for patterns and trends in data so we can make real-time predictions about future events. As researchers, evaluators and statisticians at JBS, we have been applying predictive analytic techniques for multiple industries — from health care to education to manufacturing — to forecast outcomes and target resources.

If these and other industries, such as retail and banking, can benefit from predictive analytics, so can the international development sector.

The process begins by taking stock of all available data about a given population or program. Then these data are combined and a predictive statistical model is developed to make sense of the relationships established within the data. When the model is tested and determined to accurately predict events that have already occurred, it can be used to make precise predictions about future events — all in real-time.

What do you need to successfully build a predictive model?

First, it is helpful to have a large volume of good-quality, diverse data. For example, you need data that aren’t missing a great deal of information or were poorly collected.

Second, you need statistical software that develops the predictive models and analysts who are trained in using this methodology. When little data appears to be available, we at JBS have worked with programs to recognize how to turn readily available program information into usable data and easily collect new data in seamless, efficient ways. Data is often at your fingertips!

How can predictive analytics be applied to gender issues?

The possibilities of integrating predictive analytics into programs designed to promote gender equality and the welfare and rights of women and girls are endless. Here are some examples:

For practitioners: Applying predictive analytics to promote health in girls exposed to violence

• Use individual-level data (such as demographics, health and behavioral health history, exposure to violence).

• Develop a predictive model to identify physical and behavioral health patterns in girls living in conflict regions.

• Implement a model to predict which girls are at highest risk for developing health problems that are physical (e.g. birth complications, hypertension) or behavioral (e.g. post-traumatic stress disorder, suicide and self-harm, drug addiction).

• Target at-risk girls with medical and psychosocial interventions before full-blown symptoms set in, thereby reducing the social, economic and health impacts of physical and mental illness.

For program administrators, evaluators, and planners: Applying predictive analytics to intervene in hotspots of violence against women

• Use characteristics of communities within a region (like population size, male-to-female ratio, or changes in economic, environmental, demographic and political indices).

• Develop a predictive model to anticipate outbreaks of violence toward women within various • localities.  

• Implement the model in real-time to pinpoint areas that possess a high probability of violence toward women.

• Target resources (e.g. law enforcement, public awareness materials) to potential hotspots to decrease the incidence of violence towards women and stretch dollars to focus on the areas that require the most monitoring.

For donors: Applying predictive analytics to enhance grantees’ effectiveness in working with women and girls

• Use a database of grantee characteristics (e.g. years providing services, staff size and composition, number and type of current projects).

• Develop a predictive model to identify which grantees may require additional support in a newly funded project.

• Implement model to deliver proactive training and technical assistance early in the funding cycle.

• Target delivery of T/TA to ensure success of the operation, save costs of providing more intensive T/TA down the road, and adjust future RFP criteria and/or proposal scoring criteria.

Through our experience using predictive and other advanced analytic techniques for a range of projects targeted to improve lives, we at JBS know that the efficient use of available data and real-time delivery of results make predictive analytics a natural fit for international development. Such tools harbor the potential to revolutionize how those working to empower, support and protect women and girls maximize available resources, target individuals in need, plan for the future — and, ultimately, save costs.

More simply put, predictive analytics can teach us how to innovate and how to make data work better for us and for the women and girls of the world.

Want to learn more? Check out She Builds and tweet us using #SheBuilds.

She Builds is a month-long conversation hosted by Devex in partnership with Chemonics, Creative Associates, JBS International as well as the Millennium Challenge Corp., United Nations Office for Project Services and U.K. Department for International Development.

About the authors

  • Susan Hayashi

    Susan Hayashi is a JBS vice president and offers more than 20 years of research and evaluation experience. She provides direct management and corporate oversight for projects ranging from multiyear, multimillion-dollar government contracts to rapid turnaround projects for nonprofit organizations. Many of these projects are research and evaluation projects that incorporate predictive analytics and encompass topics such as HIV/AIDS, health disparities for minority women and substance abuse.
  • Manu Singh

    Manu Singh has a doctoral degree in clinical psychology and 14 years of experience directing research and evaluation projects with particular expertise in helping clients maximize the use of data for decision making. Areas of expertise include data analytics, predictive analytics, program performance monitoring and evaluation, conducting and coordinating randomized controlled clinical trials, and secondary data analysis.
  • Radha Gholkar

    Radha Gholkar is a licensed clinical psychologist with 10 years of experience in research and evaluation including data analytics, predictive analytics, randomized controlled clinical trials, national cross-site evaluations and program development, evaluation and improvement in community mental health and correctional centers. She is recognized for providing results in user-friendly formats to support implementation of findings.