Social system engineers gain insights from survey data

At the University of Hyogo’s Graduate School of Applied Informatics, students learn how to apply social systems engineering strategies to local policy and economics. Those students often enroll in a course called Policy and Management Informatics.

This innovative course, and others like it at the Graduate School of Applied Informatics, teach future engineers and social scientists not only to understand advanced policy and economic concepts, but to wield the tools that will help them quantify the trends they observe.

Social systems engineering is an interdisciplinary field that combines economics, management, informatics, environmental studies and systems engineering using a social design and management approach to the improvement of society at large. Social systems engineers use statistical analysis to gain insights from survey data.

It goes without saying that the array of applications in this field is very wide. And students bring to the course a range of expertise – from local land management to health care and urban planning. Regardless of students’ intellectual diversity, however, they all rely on JMP^® to run their analyses.

Teaching research methodology by example

How do you teach research methodologies and analysis? By encouraging students to dive right in. And Professor Masahiro Arima leads by example.

In the lab, Professor Arima and his students investigate the influence of ICT on local commerce and governance, as well as the role played by information systems within organizational management. To do so, they look at scores of statistical data and the results of resident opinion and business surveys. The scope of this research is vast – it even takes into account the effect of ICT on the demand for culture and arts.

"Whereas civil engineering changes the world physically, social engineering is a field that seeks to make society better by influencing the people and organizations that make up communities. Survey-based opinion polls are extremely important," Arima says.

The University of Hyogo introduces students to analytics through data visualization

According to Arima, software like SAS^® provides advanced statistical analysis through programming, but JMP provides an added advantage: Researchers can analyze results with just a few clicks of the mouse.

To prepare his class to address questions that arise in the field, Arima covers topics such as the use of government statistics, design of survey questionnaires, sampling methods, techniques and problems related to carrying out surveys, data cleansing, data conversion and support for big data.

"The data gathered from surveys tends to be dirty, " he says. "Mistakes occur during the re-entry of data collected by paper or fax, and data with questionable legitimacy gets mixed in. But if you can collect clean data, JMP ensures high-quality analysis."

In 2006, Arima’s lab carried out a survey to quantitatively evaluate the needs of citizens in Itami City in regard to a local portal site. What kinds of information were residents seeking? In which formats did they want to view results? And through which organizations should information be provided?

By applying a conjoint analysis, Arima and his students clarified the preference structure for managing organization/information provided/cost allocation. In concrete terms, the team presented 12 profiles generated from three attributes (e.g., information content, use fee, managing organization) and four levels. They also used three distinct types of answer methods (e.g., yes/no evaluation, five-level rating evaluation, ordered evaluation).

However, the presumed results indicated no significant deviation between the three methodologies. "We learned that instead of the cumbersome rank logit model for ranked evaluations, even with a yes/no evaluation which allows respondents to answer intuitively, it was possible to get data that did not introduce significant deviation," Arima notes.

Arima is currently exploring the use of stated and revealed preference data in big data analytics. "Even in this era of big data, a double-pronged approach is necessary. And more thought must be given to questionnaire design and the cleaning of collected data. We will continue to research how conjoint analysis can help achieve a deeper understanding of consumer behavior and resident preference in order to assist companies and local governments in structural improvements."

Improving local policy

As a regional institution, the University of Hyogo has a community mandate—it strives to provide value not only to its students but to its community neighbors as well.  And those two goals can be achieved in tandem, especially when it comes to social systems engineering. The University’s goal is to produce analytically-minded graduates who will contribute to their communities and to society at large.

Tailoring local policy to meet the needs of the community is a theme that is echoed in his research – and that of his students. Take for example their ongoing work promoting public safety. Arima is now working on the development of a smartphone app that can determine whether a neighborhood is dangerous using the phone’s GPS function and referring to a variety of hazard maps for that location.

In the near future, he plans on running a survey of local hazard maps across Japan. And it seems likely, his students say, that JMP will have a big role to play in that project as well.