need today
Nokcool
DQWeek2Statisticsresponses.docxWednesdaySep 23 at 4:31pm
When comparing graphical analysis to quantitative analysis, it is important to think about the sample and population components of data. Quantitative analysis, by means of the mean, median, standard deviation and other summary statistics can help data scientists gauge a data sample by using one figure to represent many data points. For example, the mean tells us the average observation and what to expect from an additional data point. On the other hand, graphical analysis tends to show the population of data more succinctly and efficiently. Different graphs such as histograms and frequency distributions tell us the occurrence of data points across the entire population of data. We are able to spot population trends more easily. This can work in conjunction with quantitative data of a sample to indicate what the current data set looks like while allowing us to predict to some degree what new or additional values may be.
Graphical analysis is important to research as it is typically inclusive of larger populations of data. It also provides “quick pictures of the distributions” of data in an easy-to-digest and readable form. In real world tests, there are usually multitudes of samples to create a single data set in order to ensure efficacy and accuracy of any statistical test. Guidelines for presenting graphs should rely on their honestly and transparency to show strong data models. This includes a range of values on the axes which are within the bounds of the existing dataset, keeping the representation of objects concise, and using proper scales of data. If the axes are too narrow or feature gaps that are too wide in the data, trends may be glossed over or not completely represented. When objects are represented in misleading form such as icons that do not match the objects being measured, insights can be misconstrued. Lastly, the scales on which a graph is built tell us about the range of data outcomes, and with too large a scale, data can be muted.
References:
Lind, D. A., Marchal, W. G., & Wathen, S. A. (2017). Statistical techniques in business and economics. (17th ed.). Boston. McGraw-Hill/Irwin
2. Cameron Izzi
WednesdaySep 23 at 9:22pm
Graphical analysis is an import tool because it will allow the user of the data to see things that may not be apparent from a standard view of the data. Lind et al. says "Descriptive statistics organize data to show the general pattern of the data, to identify where values tend to concentrate, and to expose extreme or unusual data values" (Chapter 2). It is a very powerful tool, definitely when used to present something about the data to an audience as it can provide and easy visual way for someone who is unfamiliar with the data to get a quick understanding of what the data means. Some examples of things that might not be apparent within the data from a standard view is something like a change in trend over time. This could be how a user is labeling data over years, and how that label changes from a long span of time. If the data is presented using graphical analysis, plotting the data over all its time, it might become very clear how a label is changing. This is an actual use-case that I have encounter in my daily work, which when presented to the customer open the door to an extremely useful conversation.
With that being said, graphical analysis can be misleading. It can be very misleading if the data is not present properly in this form. If labels, sizing, scaling, or even color schemes are off it can give the wrong impression to an audience leading them to be mislead. It is extremely important to make sure what is being graphical represented is being done correctly and visual scaled right for the message.
References:
Lind, D. A., Marchal, W. G., & Wathen, S. A. (2017). Statistical techniques in business and economics. (17th ed.). Boston. McGraw-Hill/Irwin
WednesdaySep 23 at 5:21pm
While generally widespread knowledge already, physical activity has been shown to have very beneficial effects for both physical and mental health of individuals. I examined a research study discussing the effects of physical activity in Sweden in the 21st century and the related health effects. The study was somewhat biased in viewing sports as the primary arbiter of physical activity though other forms of movement and exercise were considered for the results. The authors propose that our daily lives are becoming less physically active, while organized exercise and training increases. Charts are used throughout to highlight important data findings as well as to propose guidelines on positive benefits of physical activity meant for specific demographic groups. For example, healthy activity guidelines and thresholds are listed for the following age groups, children and youth, 6-17 years; adults, ages 18-64; and the elderly. On the other hand, various graphs show the data outcomes for reported levels of physical activity. A featured pictograph uses different sized circles to highlight reported fatigue while a red or blue shading denotes the level of “good” or “bad” health reported in the sample. Other charts and tables feature other qualitative data such as the health risks mitigated by exercise and the associated decrease in risk as well as the general lower cause of death summarized in a bar chart. Overall, this combination of graphical analysis does provide more detailed qualitative results in text-based formats that allow for more context and specific information regarding risk prevention as a result of exercise. The quantitative graphs provide more of a glance at the research samples studied and give more insight to the level of effects reported regarding exercise. It highlights the growing trends of exercise and sport participation by youth and how those traits may carry into adulthood to help with limiting many ostensibly growing issues around obesity and other health complications from sedentary lifestyles.
References:
Malm, C., Jakobsson, J., & Isaksson, A. (2019). Physical Activity and Sports-Real Health Benefits: A Review with Insight into the Public Health of Sweden. Sports (Basel, Switzerland), 7(5), 127. https://doi.org/10.3390/sports7050127
4. Cameron Izzi
ThursdaySep 24 at 7:14am
I did research on how a group of researchers used graphical analysis to exploit data from network attacks. They used graphical analysis techniques to review capture packets from a network to try to detect if the network is being attacked from a malicious source. Aryeh et al. said "The capturing of network traffic in the form of PCAP and converting to CSV format or DataFrame can sometimes be tedious and intimidating" (Sec. C, para 6). They complied the collected data into a tabular format, the visualize the data using graphical analysis techniques to get a full understanding of what is going on in the data. They used pie charts to look at the ratio of IP address hitting the network over time, and bar charts to analyze the amount of times an IP address hits the network in milliseconds. They were able to concluded that some of the thought be malicious payload of IP address were most likely errors or misconfigurations of an access point, not actually malicious plays. Aryeh et al. said that these techniques made it an "effortless to glean" to analyze and use the data.
I found this research very interesting since I am a Machine Learning Engineer and lead a team of data scientist in my daily job. This was published in 2020 and I found they techniques they used to be out-dated, as there a plenty of off the shelf tools to do anomaly detection over a graphical representation of your graph. Also ML is huge application to network malicious detection and their are many techniques out there that do it well, that were not explored in this paper. But they were able to still prove the old ways of doing things are still a valid way of using your data, and prove effective which is awesome to me.
References
Aryeh, F. L., Alese, B. K., & Olasehinde, O. (2020). Graphical analysis of captured network packets for detection of suspicious network nodes. 2020 International Conference on Cyber Situational Awareness, Data Analytics and Assessment (CyberSA), Cyber Situational Awareness, Data Analytics and Assessment (CyberSA), 1–5. https://doi-org.proxy-library.ashford.edu/10.1109/CyberSA49311.2020.9139672
