Experience — Iris Analytics Group

Dr. Schabla combines his interest in group identity and industrial organization with his economics and statistical modeling background to solve complex problems. Through his extensive research into group identity, Dr. Schabla became an expert in technical communications, mathematical model development, and multi-disciplinary data science. Today, these expertise can be applied to many fields that require data analytics.

During his time in the Navy, Joshua trained in the operation of nuclear power plants onboard modern naval vessels. He went on to maintain and operate nuclear power plants onboard the USS Enterprise.

Iris Analytics Group is a veteran and Native American owned, small disadvantaged business headquartered in Denver, CO.

Skills & Expertise

Multi-disciplinary Data Science

Unique ability to understand data challenges across a wide range of sectors, identify data needs, obtain the necessary data, analyze the results, and report the findings. One example includes evaluation of Federal Emergency Management Agency historical storm data to estimate the number of households likely to benefit from a solar home system solution in the event of a power outage from a hurricane. Another example is work with the Air Force PFOS program including epidemiological studies on the negative health effects of perfluorinated compounds and applying statistical techniques to hydrogeological models in order to provide contaminant risk evaluation.

Technical Communications

Able to communicate complex statistical models to a non-technical audience in easy to understand and visually impactful ways. Examples include creating new styles of graphs to convert complex, non-linear, discrete probability predictions of identity formation so that these predictions could be easily understood by non-technical audiences. For my work with the Air Force, I create easy to digest line graphs to demonstrate the different ways treating data below detection thresholds can impact predictions. These graphs combine results that anyone can understand intuitively with a highly technical model to demonstrate how the technique works and why it’s the right way to solve the problem in an easy to digest way.

Staff Training and Development

Equipped to train and mentor junior and mid-level staff in data science, mathematical modeling, and statistical testing methods. Extensive teaching experience includes math tutoring, serving as a Teaching Assistant, and serving as a Graduate Instructor.

Mathematical Model Development

Extensive experience in transforming non-mathematical questions into mathematical models that allow for precise measurement and testing of initially vague ideas. Examples include providing a mathematical foundation for psychological theories on group identity and converting consumer preferences into quantifiable demand functions.

Software

STATA, EViews, R, Python, SQL, Matlab, Mathematica, GAMS, Linux

Data Analysis

Quantitative methods, data mining, machine learning techniques, business intelligence, data structure, discrete choice regressions, regression analysis, data visualization, data science research methods, research data management, statistical computing methods, and experimental design & analysis

Econometrics

Toolkit includes discrete-choice econometrics including probit, logit, rank ordered, mixed effects, and nested models as well the standard econometrics toolkit including standard OLS, selection, fixed and random effect, difference-in-difference, regression discontinuity, and IV models.

Project Experience

Environmental

Currently working with the Air Force perfluorinated compounds program. Specific responsibilities include the organization, validation, and analysis of millions of samples. Ongoing projects include statistical predictions of perfluorinated groundwater contamination levels below detection thresholds, epidemiological studies aimed at setting exposure limits, chemical fingerprint analysis to determine sources of contamination, and overall data management for a program involving millions of observations collected over more than a decade by dozens of organizations without proper data management practices being utilized.

Startup Support

Responsible for data acquisition, data mining, data analysis, and reporting of results for a clean energy startup which seeks to apply proven solar home system technology to new applications in the USA. Specific applications include the identification of remote, rural households in tribal communities who require electricity access and determination of the market potential for deploying these systems as a disaster response intervention to mitigate against prolong power outages due to hurricanes.

Publications

The history of social identity theory applied to economics traces back to the seminal work of Akerlof and Kranton, 2000. Since that time, many papers have explored the importance of this theory in experimental and theoretical settings. In this thesis I explain the need for a new model of social identity theory. Two significant problems of research into social identity theory have been an unclear definition of what an identity is and a conflation of identity with groups. I clarify these issues by providing a clear definition of identity and clarify the distinction between public and private identities.

Next, I show how identities are formed and manipulated using a novel source of micro level data. Using data from the 2004 General Social Survey and Wisconsin Ads Project, I develop a model which demonstrates how public identities are influenced by long-term and short-term priming. There are different effects between and within groups. In general, I find increasing exposure effects with increasing priming. Further, I find that individuals who exhibit behavior's indicative of possessing a stronger identity experience little impacts from external priming. Those who select an identity but do not exhibit behavior's associated with that identity are what drive the exposure effects. These results suggest that identities can be manipulated to alter individual's choices.