Introduction
A histogram is a representation of the distribution of numerical data, typically in the form of a bar chart, with “bins” or “buckets” on the x-axis and some numeric measure on the y-axis. The height of each bar corresponds to the frequency data points within the range defined by each bin.
Histograms can be used to identify patterns in the data, as well as statistical measures describing the data such as mean, median, mode, skewness, and so on. From these patterns, actionable decisions can be made (Goetsch & Davis, 2021, p. 246-251).
The following example describes how a histogram of website traffic would be useful for the owners and operators of a recipe website. The conclusion shows how traffic analysis useful not only to other types of websites but also to brick-and-mortar stores.
Working Example: A Recipe Website
An application of histograms at a software or web hosting company is as follows: the “bins” on the x-axis will be the hours of the day (so there will be 24 bins), and the height of the bars will be the number of visitors this web hosting company receives per hour.
In the following, we consider a company that operates a recipe website such as allrecipes.com or epicurious.com. The histogram would show the number of visitors (users) in each hour of the day. There would have to be adjustments for different time zones, of course.
The histogram would show that there are traffic spikes in the number of visitors at 4AM, 11AM, and 5PM, as well as a spike at around 8PM. Why these spikes? What actions can be taken based on the times of these spikes?
There are three types of employees which would use this histogram: managers, salespeople, and network engineers.
Managers would look at the histogram and hypothesize that the peak at 8PM represents visitors looking for recipes for the next day’s meals. The 4AM spike corresponds to visitors following recipes while preparing breakfast, 11AM for lunch, and 5PM for dinner.
These hypotheses can be verified by looking at the kinds of recipes the visitor uses for each meal. For example, recipes for steak and eggs should be requested for the breakfast peak but not the lunch or dinner peaks. This type of analysis is not easy, as the website uses a database to serve visitors’ needs, but it is possible (Ihm & Pai, 2011). The manager can then tailor the content to correspond to visitors’ needs.
The salesperson would use the frequency of visits to determine advertising rates: an advertiser would be charged more for ads shown during one of the peaks and would be charged less during the off-hours (Pande et al, March 2014).
A network engineer would be interested in this histogram for two reasons. One would be to schedule the right number of computers to handle the expected number of visitors at each hour: more computers would be needed during peak hours, and fewer computers would be needed during the off-hours. The decision can also be made using artificial intelligence (Wong et al, April 2005), but human-made decisions are almost always sufficient.
The other reason a network engineer would use this histogram is to determine the best time to make modifications to the website and corresponding databases. The rule is: only make modifications during off-hours. That way, if the modification introduces any breaking changes, the number of visitors affected by the change will be minimal, and there will be time to correct those changes before there is a rush.
Conclusion
As this example shows, a histogram of traffic data is useful to owners and operators of a website. It requires different types of employees to analyze the data shown in the histogram and to act based on that data. The advertising rates decided by salespeople can optimize company revenue, and the actions taken by managers and network engineers serve to deliver a more valuable website to the visitors, thus increasing customer satisfaction.
This type of analysis is not limited just to recipe websites, it applies to the traffic experienced by any website. Also, it is not limited to just websites: brick-and-mortar stores would track and analyze the number of customers entering the store and adjust the number of customer service representatives accordingly.
References
Goetsch, D. L. & Davis, S. B. (2021). Quality management for organizational excellence: Introduction to total quality (9th ed.). Pearson.
Ihm, S. & Pai, V. (2011). Towards understanding modern web traffic. IMC '11: Proceedings of the 2011 ACM SIGCOMM conference on Internet measurement conference, 295-312. https://doi.org/10.1145/2068816.2068845
Pande, P. et al. (March 2014). A study of web traffic analysis. International Journal of Computer Science and Mobile Computing 3(3), 900-907. https://www.academia.edu/6585628/A_Study_of_Web_Traffic_Analysis
Wong, X. et al. (April 2005). Intelligent web traffic mining and analysis. Journal of Network and Computer Applications 28(2), 147-165. https://doi.org/10.1016/j.jnca.2004.01.006
No comments:
Post a Comment