Conference paper · 2015
Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study
2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy, pp. 489–496, IEEE · Published
Summary
Vajjhala et al. (2015) applied simple correspondence analysis in SPSS 21 to an open big dataset of 125,087 global terrorism records (1970–2013) in a positivist exploratory case study, relating nine attack methods to 13 world regions. The first two dimensions captured 94.5% of total inertia (0.687 and 0.258), justifying a two-dimensional symmetric plot, while the overall association between attack type and region was low (0.21). The authors present this as an accessible method for visualizing open big data and aim to generalize the methodology rather than the specific results.
Practical implication. Decision makers can use correspondence analysis in common statistical software to turn qualitative big data into visual relationship maps, without specialist big data infrastructure.
- The Problem
- Decision-makers need to find hidden relationships in large qualitative open datasets without specialist big data infrastructure.
- The Methodology
- Positivist exploratory case study applying simple correspondence analysis in SPSS 21 to an open big dataset of 125,087 global terrorism records (1970–2013), relating nine attack methods to 13 world regions.
- The Core Finding
- Simple correspondence analysis in standard statistical software, applied to 125,087 global terrorism records (1970–2013), related nine attack methods to 13 world regions with two dimensions capturing 94.5% of inertia — demonstrating an accessible method for visualizing open big data.
- The Citation
- Vajjhala, N. R., Strang, K. D., & Sun, Z. (2015). Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study. In 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy (pp. 489–496). IEEE. https://doi.org/10.1109/FiCloud.2015.15
The summary, key findings, methods, implications and relevance notes on this page are this website’s own description of the paper, written from the published abstract and text. The official abstract and citation details are given further down.
Key findings
- Finding 1. The study analyzed a global terrorism open meta big dataset of 125,087 records with 133 fields (16,636,571 data points) covering 209 countries grouped into 13 regions from 1970 to 2013.
- Finding 2. The terrorism big data study coded attack method into nine nominal types: assassinations, armed assaults, bombing/explosions, hijacking, barricading, kidnapping, infrastructure, unarmed assaults and unknown/mixed.
- Finding 3. Correspondence analysis of attack method by region found that the first dimension captured 0.687 and the second 0.258 of total inertia, together 94.5% (0.945).
- Finding 4. The third correspondence analysis dimension added only 3.4% (0.034) of inertia, which the authors used to justify a two-dimensional symmetric plot of terrorist attack type by region.
- Finding 5. The estimated overall association (correlation) between terrorist attack type and geographic region in the global terrorism big data was 0.21, described as low.
- Finding 6. The study identified armed assault, bombs and assassinations as the most lethal terrorist attack methods of concern to organizations.
- Finding 7. The authors argue that correspondence analysis is preferable to cluster analysis or chi-square tests for nominal big data because it estimates the strength and direction of associations among factors without losing attribute detail.
Source: Vajjhala et al. (2015), 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy, pp. 489–496, IEEE. DOI: 10.1109/FiCloud.2015.15
The study in detail
What question does this paper answer?
How could qualitative big data be collected and analyzed, using readily available statistical software, to identify hidden factor relationships that support strategic decision making?
What did the study find, in detail?
Applying simple correspondence analysis in SPSS 21 to an open meta big dataset of 125,087 global terrorism records (133 fields, 16,636,571 data points, 1970–2013), the authors related nine terrorist attack methods to 13 world regions. The first two dimensions captured 94.5% of total inertia (0.687 and 0.258), justifying a two-dimensional symmetric plot, while the overall association between attack type and region was low (0.21).
Why does it matter?
The study shows that qualitative big data can be analyzed with desktop statistical software to reveal hidden relationships that business managers could use when selecting marketing partners and supply chain providers or deciding to scale down operations in risky areas. The authors aim to generalize the methodology rather than the specific results, and call for more empirical big data analytics papers and a methodology guidebook for practitioners and researchers.
Methods and data
| Research question | How can qualitative big data be collected and analyzed to identify factor relationships that support strategic decision making? |
|---|---|
| Design | Positivist, exploratory empirical case study using descriptive non-parametric statistics |
| Data | Global terrorism open meta big dataset: 125,087 records, 133 fields, 16,636,571 data points, 209 countries in 13 regions, 1970–2013 |
| Methods | Simple correspondence analysis (symmetrical normalization, Euclidean distance) in SPSS version 21 of attack method by region |
| Main result | Two dimensions explained 94.5% of inertia (0.687 + 0.258); overall attack-type–region association was 0.21 (low) |
| Implication | Correspondence analysis with common software can turn qualitative big data into visual relationship maps for decision makers |
| Citation | Vajjhala et al. (2015) · DOI 10.1109/FiCloud.2015.15 |
Definitions and key terms
- Correspondence analysis
- A multivariate exploratory technique that estimates the interdependence (inertia) between categorical variables in a contingency table and plots them as distances in a low-dimensional map. — as used in Vajjhala et al. (2015)
- Open big data
- Very large datasets made publicly available, here a meta dataset summarizing terrorist activity as keyword frequencies from public news articles. — as used in Vajjhala et al. (2015)
- Inertia
- In correspondence analysis, the measure of association between row and column factors, whose proportion accounted for by each dimension indicates how much of the relationship a plot captures. — as used in Vajjhala et al. (2015)
Abstract
This study addressed the literature gap, identified by other researchers, that there are too few examples of applied empirical open big data analytics. Using correspondence analysis as a big data analytical technique, this study demonstrates how qualitative big data type could be analyzed to identify hidden factor relationships that may assist strategic decision making. We use a 64MB open meta big dataset developed by summarizing terrorist activity as keyword frequencies collected from trillions of public news articles published during a 43 year period from 1970-2013 and readily available statistical software, SPSS, to visually summarize the findings on a global terrorism big dataset. The approach in this paper might facilitate the research and development of open big data, big data analytics against global terrorism.
Abstract as published in 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy.
Keywords: open big data analytics; correspondence analysis; global terrorism open meta big dataset; statistical modeling; data mining; data warehouse
Practical implications
- Business managers could use hidden relationships revealed by analyzing qualitative big data with desktop statistical software when selecting marketing partners and supply chain providers or deciding to scale down operations in risky areas.
- Researchers analyzing nominal big data may prefer correspondence analysis to cluster analysis or chi-square tests, because the authors argue it estimates the strength and direction of associations among factors without losing attribute detail.
- The authors aim to generalize the methodology rather than the specific results, and call for more empirical big data analytics papers and a methodology guidebook for practitioners and researchers.
Drawn from Vajjhala et al. (2015), 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy, pp. 489–496, IEEE.
When this research may be relevant
This article may be relevant to researchers analysing the Global Terrorism Database or other open big data with descriptive multivariate statistics, applying correspondence analysis to nominal big data, or visualising qualitative big data for strategic decision-making without specialised big data infrastructure.
Research topics addressed: big data analytics; open big data; global terrorism; Global Terrorism Database; correspondence analysis; statistical modeling; data visualization; SPSS; qualitative big data; nominal data; data mining; data warehouse; strategic decision making; terrorism risk
How to cite
- APA 7
- Vajjhala, N. R., Strang, K. D., & Sun, Z. (2015). Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study. In 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy (pp. 489–496). IEEE. https://doi.org/10.1109/FiCloud.2015.15
- MLA 9
- Vajjhala, Narasimha Rao, et al. “Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study.” 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy, IEEE, 2015, pp. 489–96. https://doi.org/10.1109/FiCloud.2015.15.
- IEEE
- N. R. Vajjhala, K. D. Strang, and Z. Sun, “Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study,” in 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy, 2015, pp. 489–496, doi: 10.1109/FiCloud.2015.15.
BibTeX
@inproceedings{vajjhala2015open,
title = {Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study},
author = {Vajjhala, Narasimha Rao and Strang, Kenneth David and Sun, Zhaohao},
booktitle = {2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy},
pages = {489--496},
year = {2015},
publisher = {IEEE},
doi = {10.1109/FiCloud.2015.15},
url = {https://doi.org/10.1109/FiCloud.2015.15}
}RIS
TY - CPAPER TI - Statistical Modeling and Visualizing Open Big Data Using a Terrorism Case Study AU - Vajjhala, Narasimha Rao AU - Strang, Kenneth David AU - Sun, Zhaohao T2 - 2015 3rd International Conference on Future Internet of Things and Cloud (FiCloud), Rome, Italy PY - 2015 DA - 2015/08 SP - 489 EP - 496 PB - IEEE DO - 10.1109/FiCloud.2015.15 UR - https://www.narasimharao.net/research/open-big-data-correspondence-analysis-global-terrorism/ LA - en ER -
Links and files
Related research
Related publications on machine learning and data science applications
The survey organizes visual data mining techniques, tools and collaborative filtering approaches, and concludes that combining visual data mining with collaborative filtering has led to substantial improvement and can alleviate the shortcomings of traditional collaborative filtering.
Across 92 primary studies (2016–2025), Diffusion Transformers and Gaussian Splatting account for about 36% of studies dated 2024 or later while GAN use has fallen below 3%; training data are demographically skewed and under-documented, and the dominant metrics (PSNR, SSIM, SyncNet) do not measure what deployment requires.
Fine-tuned DINOv2 and EfficientNet-B0 performed best on nine-class DeepWeeds data, but differences among several high-performing models were not consistently statistically significant across random seeds.
On ten years of daily data for Sofia, Berlin and Tokyo, linear regression matched random forest for temperature at a fraction of the run time and clearly outperformed it for humidity (R² 0.940–0.964), while decision trees performed worst.
A prefetching approach that builds an online navigation graph from server logs and clusters navigation behaviour with WEKA outperformed conventional prefetching techniques on domain client-group logs, particularly once server idle time allowed prefetching across all user predictions.