Sector Analytics

Sector Analytics

It is a basic approach of doing research for economic sectors what is in the description below about bankruptcies in a socio-economic sector.

TitleUsing  Data Science in identifying the factors for Strategic Risk assessment in business interruption/bankruptcy prediction      
KeywordsStrategic risk, bankruptcy, data science, artificial intelligence, data analytics, business interruption
AbstractFor many years the process of predicting business interruption / bankruptcy several factors and methodologies have been used, in this study the main purpose is to review how the conditions of certain factors are affecting the capacity to perform before, in and after crisis. The capacity to and appropriate performance is affecting the probability of business interruption or bankruptcy. Machine learning based techniques are used for the prediction of probability of losses (default severity)
DescriptionBankruptcy and  Business interruption are more often in a society that is under uncertainty. Multiple risk factors can create chain reactions in organizations. This research is based on the comparison of organizations that went to bankruptcy or stop operations vs. the ones that were with a good standing performance.  bankruptcy cases that occurred in the years before, in and after crisis. At the same time, taking into account the characteristics of the financial/epidemic crisis, failure detection models and their evaluation models are developed as a tool for organizations to deal with short-term, medium-term and long-term risks.  In particular, the problem will be tackled by classification of organizations and grouping them by certain factors (variable clustering) and supervise/reinforcement learning such as Deep learning and ANN using multi-layer models that support the analysis of the data and diverse socio-economic circumstances. Variables and analysis by Industry, world regions and country risk are very important topics when analyzing interruption/ bankruptcy risk. Due to this fact, thorough consideration should be given to it when modeling these kind of risks. Three approaches are common ground in the industry. The first approach or the ideal one is to have one model per combination of country and industry. The second is to fit several models per industry sector or per geographic area. The third is to introduce categorical variables to account for industry and country specific variations.  There is process to review the data available to the level of granularity to avoid complex unbalanced data to work with. Having different models per combination of industry and country is not feasible, given the amount of claims information that would be needed per sector. Information would be unavailable in many of the combinations or in some cases results would not be robust and thus lack any meaning.
GoalsTo Identify factors influencing the probability of defaultTo reflect about the use of analytic tools and data requirementsTo think about the solutions to problems that are not perfectly defined and maintain ambiguity in the analysisTo identify the pros and cons within the simulation approach in a real problemTo manage the complexity of data and probability distribution selectionTo structure simulation and optimization models that can support decisionsTo analyze results under different assumptionsTo take into consideration and describe how strategic risk can be analyzed
Methodology-Gather and organize data, internal and external
-Prepare a report of Exploratory Data Analysis (EDA) where key points are identified
-Create the first approach of the internal model
-Create the first approach of the external model
-Create a full model and first observations/conclusions
-Perform sensitivity analysis and stress testing
-Present final report  
Questions to answerThe core concepts to use are the following: Strategic risk and Evaluating and monitoring risk  
-What are the factors that have been studied in bankruptcy? Performance and risk indicators
-What are the metrics in bankruptcy? Variation of results
-How are the results varying before, in, and after the crisis? Building intelligence,        Data Analytics Use          
-What are the methods in data analytics used for tackling the problem? Developed solutions      
-How are the solutions implemented?   Data Science and Digitalization                  
-How are organizations using data science? Digital transformation Initiatives 
-How are organizations digitalizing? Models used for tackling the problem
-How has the digital transformation been a factor to impulse changes? Strategy techniques and technologies, Convert data into actions             
-What are lessons learned in the changes? Errors in assessment     
-How were the predictions vs. reality? Errors in implementation            
-What are the issues in implementing strategies? Interaction Human and Machine, Ethical issues
-In which ways are humans and machines solving the problems? Operational risks        
-How have the work been complemented by human and machines? who is doing what?
-What are the companies doing to deal with the interaction?
Expected outcomes-Literature review
-Data base organization
-First test of models
-Report of first observations
References Formal peer review bibliography