Clickstream Data Analysis Research Articles

Building a Data Webhouse to support clickstream data analysis

Building a Data Webhouse to support clickstream data analysis

How mobile users respond to pre-roll skippable advertising differently than personal computer (PC) users: an analysis of individual-level clickstream data

Mobile devices have evolved as a major channel for online video consumption. With the growth of online video platforms, various ad formats have been experimented to increase advertising effectiveness. One such attempt is pre-roll skippable ads that allow users to skip an ad after watching it for a few seconds. Despite pre-roll ads providing better user experience, their high skipping rate is alarming to advertisers and brands. Recent studies revealed what factors influence pre-roll ad acceptance, yet little has been known regarding the role of device type. This study investigates how mobile users respond to pre-roll skippable ads differently than PC users, focusing on individual and contextual factors of video consumption. Using individual-level clickstream data obtained from a major online video platform in South Korea, this study found that mobile users tend to skip pre-roll ads more often than PC users, but the degree differs by individual and contextual factors.

Measuring Students’ Engagement with Digital Interactive Textbooks by Analyzing Clickstream Data

This paper provides an overview of my contributions to a project to measure and predict student’s mental workload when using digital interactive textbooks. The current work focuses on analysis of clickstream data from the textbook in search of viewing patterns among students. It was found that students typically fit one of three viewing patterns. These patterns can be used in further research to inform creation of new interactive texts for improved student success.

Segmenting Clinicians' Usage Patterns of a Digital Health Tool in Resource-Limited Settings: Clickstream Data Analysis and Survey Study.

BackgroundEvidence-based digital health tools allow clinicians to keep up with the expanding medical literature and provide safer and more accurate care. Understanding users’ online behavior in low-resource settings can inform programs that encourage the use of such tools. Our program collaborates with digital tool providers, including UpToDate, to facilitate free subscriptions for clinicians serving in low-resource settings globally.ObjectiveWe aimed to define segments of clinicians based on their usage patterns of UpToDate, describe the demographics of those segments, and relate the segments to self-reported professional climate measures.MethodsWe collected 12 months of clickstream data (a record of users’ clicks within the tool) as well as repeated surveys. We calculated the total number of sessions, time spent online, type of activity (navigating, reading, or account management), calendar period of use, percentage of days active online, and minutes of use per active day. We defined behavioral segments based on the distributions of these statistics and related them to survey data.ResultsWe enrolled 1681 clinicians from 75 countries over a 9-week period. We based the following five behavioral segments on the length and intensity of use: short-term, light users (420/1681, 25%); short-term, heavy users (252/1681, 15%); long-term, heavy users (403/1681, 24%); long-term, light users (370/1681, 22%); and never-users (252/1681, 15%). Users spent a median of 5 hours using the tool over the year. On days when users logged on, they spent a median of 4.4 minutes online and an average of 71% of their time reading medical content as opposed to navigating or managing their account. Over half (773/1432, 54%) of the users actively used the tool for 48 weeks or more during the 52-week study period. The distribution of segments varied by age, with lighter and less use among those aged 35 years or older compared to that among younger users. The speciality of medicine had the heaviest use, and emergency medicine had the lightest use. Segments varied strongly by geographic region. As for professional climate, most respondents (1429/1681, 85%) reported that clinicians in their area would view the use of a online tool positively, and compared to those who reported other views, these respondents were less likely to be never-users (286/1681, 17% vs 387/1681, 23%) and more likely to be long-term users (655/1681, 39% vs 370/1681, 22%).ConclusionsWe believe that these behavioral segments can help inform the implementation of digital health tools, identify users who may need assistance, tailor training and messaging for users, and support research on digital health efforts. Methods for combining clickstream data with demographic and survey data have the potential to inform global health implementation. Our forthcoming analysis will use these methods to better elucidate what drives digital health tool use.

Predicting Analysis of User’s Interest from Web Log Data in e-Commerce using Classification Algorithms

The accelerated development of e-commerce has been a concern for business people. Business people should be able to gain customer interest in a variety of ways so that their companies can compete with others. Analyzing click-flow data will help organizations or firms assess customer loyalty, provide advertising privileges, and develop marketing strategies through user interests. By understanding consumer preferences, clickstream data analysis may be used to determine who is participating, assist companies in evaluating customer contentment, boost productivity, and design marketing strategies. This research was performed by defining experimental user interests using Dynamic Mining and Page Interest Estimation methods. The findings of this analysis, using three algorithms at the pattern discovery page, demonstrated that the Decision Tree method excelled in both methods. It indicated that the operational performance of the Decision Tree performed well in the assessment of user interests with two different approaches. The findings of this experiment can be used as a proposal for researching the field of web usage mining, collaborating with other approaches to achieve higher accuracy values.

An Empirical Evaluation of Online Continuous Authentication and Anomaly Detection Using Mouse Clickstream Data Analysis

While the password-based authentication used in social networks, e-mail, e-commerce, and online banking is vulnerable to hackings, biometric-based continuous authentication systems have been used successfully to handle the rise in unauthorized accesses. In this study, an empirical evaluation of online continuous authentication (CA) and anomaly detection (AD) based on mouse clickstream data analysis is presented. This research started by gathering a set of online mouse-dynamics information from 20 participants by using software developed for collecting mouse information, extracting approximately 87 features from the raw dataset. In contrast to previous work, the efficiency of CA and AD was studied using different machine learning (ML) and deep learning (DL) algorithms, namely, decision tree classifier (DT), k-nearest neighbor classifier (KNN), random forest classifier (RF), and convolutional neural network classifier (CNN). User identification was determined by using three scenarios: Scenario A, a single mouse movement action; Scenario B, a single point-and-click action; and Scenario C, a set of mouse movement and point-and-click actions. The results show that each classifier is capable of distinguishing between an authentic user and a fraudulent user with a comparatively high degree of accuracy.

Evaluating Chinese government WeChat official accounts in public service delivery: A user-centered approach

WeChat official accounts have been increasingly adopted by Chinese government agencies to deliver public services, in response to the “Internet + Public Service” reformation. While previous studies depended heavily on the expert-oriented approach to evaluate the accounts, this paper presents a user-centered study based on a mixed methods research design in which an unobtrusive clickstream data analysis was complemented by a card sorting study, stakeholder interviews, and a focus group. A 2-month server log file containing 42,188,760 clickstream records was obtained from an active government WeChat official account and analyzed at the movement level, which found that the account was mainly used as a lookup tool with most services underutilized and its home portals failed to support effective wayfinding to needed services. Deficiencies in information architecture, operation strategy, and interaction design of the account were identified in the complementary studies. This study not only enriches the knowledge about social media use in the Chinese government for public service delivery, but also introduces innovative methods to generate new research insights. The findings can inform government WeChat official accounts of how to improve service quality and user experience.

The Impact of Social vs. Nonsocial Referring Channels on Online News Consumption

The digitization of news markets has created a key role for online referring channels. This research combines field and laboratory experiments and analysis of large-scale clickstream data to study the effects of social versus nonsocial referral sources on news consumption in a referred news website visit. We theorize that referrer-specific browsing modes and referrer-induced news consumption thresholds interact to impact news consumption in referred visits to an online newspaper and that news sharing motivations invoked by the referral source impact sharing behavior in these referred visits. We find that social media referrals promote directed news consumption—visits with fewer articles, shorter durations, yet higher reading completion rates—compared with nonsocial referrals. Furthermore, social referrals invoke weaker informational sharing motivations relative to nonsocial referrals, thus leading to a lower news sharing propensity relative to nonsocial referrals. The results highlight how news consumption changes when an increasing amount of traffic is referred by social media, provide insights applicable to news outlets’ strategies, and speak to ongoing debates regarding biases arising from social media’s growing importance as an avenue for news consumption.This paper was accepted by Anandhi Bharadwaj, information systems.

Detecting Academic Misconduct Using Learning Analytics

Keystroke logging and clickstream data, both emergent areas of study in the field of learning analytics, present promising alternative methods of detecting and preventing contract cheating. The current study examines whether analysis of keystroke and clickstream data can detect when a student is creating their own authentic writing or transcribing from another source. Participants were 62 university students (47 women, 15 men) who completed three writing tasks under experimental conditions: free writing, general transcription, and self-transcription. Analyses revealed that while completing the free-writing task, participants typed in shorter bursts with longer pauses and typed more slowly with more revisions compared to the two transcription tasks. Model-based clustering was able to accurately distinguish the free-writing task from the two transcription tasks based on patterns of bursts and writing speed. Overall, these results suggest that keystroke and clickstream analysis may be able to distinguish between a student writing an authentic piece of work and one transcribing a completed work. These findings signal significant implications for the detection of contract cheating.

What prompts users to click on news headlines? Evidence from unobtrusive data analysis

PurposeThe headlines of online news are created carefully to influence audience news selection today. The purpose of this paper is to investigate the relationships between news headline presentation and users’ clicking behavior.Design/methodology/approachTwo types of unobtrusive data were collected and analyzed jointly for this purpose. A two-month server log file containing 39,990,200 clickstream records was obtained from an institutional news site. A clickstream data analysis was conducted at the footprint and movement levels, which extracted 98,016 clicks received by 7,120 headlines ever displayed on the homepage. Meanwhile, the presentation of these headlines was characterized from seven dimensions, i.e. position, format, text length, use of numbers, use of punctuation marks, recency and popularity, based on the layout and content crawled from the homepage.FindingsThis study identified a series of presentation characteristics that prompted users to click on the headlines, including placing them in the central T-shaped zones, using images, increasing text length properly for greater clarity, using visually distinctive punctuation marks, and providing recency and popularity indicators.Originality/valueThe findings have valuable implications for news providers in attracting clicks to their headlines. Also, the successful application of nonreactive methods has significant implications for future user studies in both information science and journalism.

Segmentifier: Interactive Refinement of Clickstream Data

AbstractClickstream data has the potential to provide insights into e‐commerce consumer behavior, but previous techniques fall short of handling the scale and complexity of real‐world datasets because they require relatively clean and small input. We present Segmentifier, a novel visual analytics interface that supports an iterative process of refining collections of action sequences into meaningful segments. We present task and data abstractions for clickstream data analysis, leading to a high‐level model built around an iterative view‐refine‐record loop with outcomes of conclude with an answer, export segment for further analysis in downstream tools, or abandon the question for a more fruitful analysis path. Segmentifier supports fast and fluid refinement of segments through tightly coupled visual encoding and interaction with a rich set of views that show evocative derived attributes for segments, sequences, and actions in addition to underlying raw sequences. These views support fast and fluid refinement of segments through filtering and partitioning attribute ranges. Interactive visual queries on custom action sequences are aggregated according to a three‐level hierarchy. Segmentifier features a detailed glyph‐based visual history of the automatically recorded analysis process showing the provenance of each segment as an analysis path of attribute constraints. We demonstrate the effectiveness of our approach through a usage scenario with real‐world data and a case study documenting the insights gained by a corporate e‐commerce analyst.

Adapting and Adopting Open Educational Resources: An Analysis of Student Cost Savings, Use, Performance, and Perception

Open educational resources (OERs) are significant in terms of cost savings to the student, but there are questions surrounding the quality of these resources, as well as whether students prefer OER or traditional textbooks. Many of the OERs available today are often peer-reviewed, but some platforms allow the users to alter or customize the content, like OpenStax Connexions (CNX). When open content is altered or rearranged by an instructor, then it becomes essential for the instructor to determine whether those changes are beneficial or detrimental to student learning. Using the OpenStax Connexions (CNX) OER educational content repository and content management system, two customized OERs were developed and used as the only textbooks for two introductory chemistry courses at a community college in rural Oregon. The au- thor of this study examined students’ cost savings, performance, use, and perception. Student scores for the OER-only courses were compared with scores from courses taught with traditional text-books. The results of Student’s t-test suggest that there was a significant difference between scores, in favor of those taught with an OER textbook. Because of small sample sizes, Cohen’s d was also calculated and indicated that, in most cases, the effect size was not large enough to be considered significant. Although it is difficult to say that learning was improved in light of the small effect sizes, it seems reasonable to suggest that learning was not adversely affected by the adoption of customized OERs. Lastly, an analysis of clickstream data from the learning management system and data obtained from an end of course survey seem to indicate that stu- dent usage and perception of OER does not differ significantly with traditional textbooks.

A clickstream data analysis of users' behavior on a government WeChat official account

ABSTRACTThis study took a user‐oriented approach to investigating the effectiveness of a government WeChat official account in terms of service delivery. A transaction log file containing 42,188,760 clickstream records was obtained from an active account and analyzed at the footprint and movement levels. The current clickstream data analysis revealed the underutilization of the services provided by the account, the failure of the three navigational portals, and the poor completion rates of the tasks users performed on the account.

A utility-based news recommendation system

News platforms exhibit both the challenges as well as opportunities for enhancing the functionalities of recommendation systems in today's big data environment. Novel use of big data storage and programming models can improve news recommendation systems through efficient handling and analysis of clickstream data and a better understanding of users' interests. Most existing approaches to news recommendation consider users' clicks as the implicit feedback to understand user behaviors. However, “clicks” may not be an effective indicator of real user interests. We address this problem by developing a novel news recommendation system based on a news utility model. Given the new utility model, we propose a two stage news recommendation framework. The framework first generates article-level recommendation rules based on the utility model, then integrates the notion of utility and probabilistic topic models and generates topic-level recommendation rules. We argue that the proposed utility-based news recommendation system also addresses the news cold start problem which is one of the most challenging obstacles for news agencies. We evaluate the framework on a massive real dataset (two billion records) obtained from a major newspaper (i.e., The Globe and Mail) in Canada and show that it outperforms the existing methods.

A Clickstream Data Analysis of the Differences between Visiting Behaviors of Desktop and Mobile Users

Mobile devices are gaining popularity among online shoppers whose behavior has been reshaped by the changes in screen size, interface, functionality, and context of use. This study, based on a log file from a cross-border E-commerce platform, conducted a clickstream data analysis to compare desktop and mobile users' visiting behavior. The original 2,827,449 clickstream records generated over a 4-day period were cleaned and analyzed according to an established analysis framework at the footprint level. Differences are found between desktop and mobile users in the distribution of footprints, core footprints, and footprint depth. As the results show, online shoppers preferred to explore various products on mobile devices and read product details on desktops. The E-commerce mobile application (app) presented higher interactivity than the desktop and mobile websites, thus increasing both user involvement and product visibility. It enabled users to engage in the intended activities more effectively on the corresponding pages. Mobile users were further divided into iOS and Android users whose visiting behaviors were basically similar to each other, though the latter might experience slower response speed.

Predicting Purchase Behavior of Website Audiences

ABSTRACTThis paper proposes a methodological framework that extends the advantages of behavioral targeting while preserving the privacy of the individual. Instead of profiling individual users according to their general interests, we profile website audiences according to their online purchase behavior. This presents a trade-off between looser, aggregate audience profiling and deeper understanding of actual purchase behavior, the holy grail of online advertising. Our framework is based on the analysis of raw clickstream data of Web users who explicitly agreed to participate in an online audience panel. We experiment with data collected by an online analytics company, SimilarWeb, which consists of 3,463,796 records of online purchases and 1.1 billion records of Website visits. We train a multilabeled classification model on the clickstream of panel members with distinctive online purchase profiles to predict the purchase potential of the entire panel. We aggregate the individual purchase behavior profiles (both ground-truth and predicted) into purchase behavior profiles of Web domain audiences and test the resulting methodology on 3,408 Web domains, with very promising results. If privacy-related regulation tightens up in the near future, the proposed panel-based, purchase-focused ad targeting mechanism might be the panacea for online advertisers.

Exposure to positive peer sentiment about nicotine replacement therapy in an online smoking cessation community is associated with NRT use

BackgroundLittle is known about the influence of online peer interactions on health behavior change. This study examined the relationship between exposure to peer sentiment about nicotine replacement therapy (NRT) in an online social network for smoking cessation and NRT use. MethodsParticipants were 3297 current smokers who enrolled in an Internet smoking cessation program, participated in a randomized trial, and completed a 3-month follow-up. Half received free NRT as part of the trial. Automated text classification identified 27,038 posts about NRT that one or more participants were exposed to in the social network. Sentiment towards NRT was rated on Amazon Mechanical Turk. Participants' exposure to peer sentiment about NRT was determined by analysis of clickstream data. Modified Poisson regression examined self-reported use of NRT at 3-months as a function of exposure to NRT sentiment, controlling for study arm and post exposure. ResultsOne in five participants (19.3%, n = 639) were exposed to any NRT-related posts (mean exposure = 6.5 ± 14.7, mean sentiment = 5.4 ± 0.8). The association between sentiment exposure and NRT use varied by receipt of free NRT. Greater exposure to positive NRT sentiment was associated with an increased likelihood of NRT use among participants who did not receive free NRT (adjusted rate ratio 1.22, 95% CI 1.01, 1.47; p = .043), whereas no such relationship was observed among participants who did receive free NRT (p = .48). ConclusionsExposure to positive sentiment about NRT was associated with increased NRT use when smokers obtained it on their own. Highlighting user-generated content containing positive NRT sentiment may increase NRT use among treatment-seeking smokers.

The Price of Privacy

The analysis of clickstream data facilitates the understanding and prediction of customer behavior in e-commerce. Companies can leverage such data to increase revenue. For customers and website users, on the other hand, the collection of behavioral data entails privacy invasion. The objective of the paper is to shed light on the trade-off between privacy and the business value of customer information. To that end, the authors review approaches to convert clickstream data into behavioral traits, which we call clickstream features, and propose a categorization of these features according to the potential threat they pose to user privacy. The authors then examine the extent to which different categories of clickstream features facilitate predictions of online user shopping patterns and approximate the marginal utility of using more privacy adverse information in behavioral prediction models. Thus, the paper links the literature on user privacy to that on e-commerce analytics and takes a step toward an economic analysis of privacy costs and benefits. In particular, the results of empirical experimentation with large real-world e-commerce data suggest that the inclusion of short-term customer behavior based on session-related information leads to large gains in predictive accuracy and business performance, while storing and aggregating usage behavior over longer horizons has comparably less value.

The Impact of Social vs. Non-Social Referring Channels on Online News Consumption

The digitization of news markets has created a key role for online referring channels. This research combines field and lab experiments, and analysis of large-scale clickstream data, to study the effects of social versus non-social referral sources on news consumption in a referred news website visit. We propose that referring channels generate a new type of priming effect, denoted the referrer effect, as unique features of the referrer affect user behavior in a subsequent news visit. We find that social media referrals promote focused reading – visits with fewer articles, shorter durations, yet higher reading completion rates - compared to non-social referrals. Furthermore, social referrals decrease news sharing propensity, due to lower perceived novelty to peers of content discovered via social media. The results provide insights applicable to news outlets’ social media strategies, and speak to ongoing debates regarding biases arising from social media’s growing importance as an avenue for news consumption.

Unpacking the Strategies of Ten Highly Rated MOOCs: Implications for Engaging Students in Large Online Courses

Background/Context The advent of massive open online courses (MOOCs) has fueled much attention among educators. However, despite the high interest they generate, we still understand little about student engagement in these large-scale online courses. Existing studies tend to focus on how MOOCs affect higher education institutions in terms of faculty identity, workload, responsibilities, and policy. Other studies have mostly employed clickstream data analysis to predict student dropout or completion. Although studies such as these are useful, they fall short of explaining the reasons why participants find the activities or course engaging. Research Questions Unlike many previous studies, this study seeks to uncover what factors related to MOOC pedagogy or to the individual instructor may encourage or discourage student engagement. This study explores the following questions: What elements related to the course design or the teaching staff did students find enjoyable, beneficial in helping them learn the materials, or motivate them to participate in the activities? What elements did students wish to improve? What elements related to the course design or teaching staff did students find frustrating? Participants The sample consisted of 4,466 learners who participated in one or more of 10 highly rated MOOCs. Highly rated MOOCs were sampled because they were likely to exemplify good practice or teaching strategies. Research Design Qualitative research methods were used in this study. More specifically, detailed observations of the 10 MOOCs’ course features and grounded analyses of the 4,466 learners’ course review data were conducted. Findings Findings suggest six key factors that can engage online students and nine reasons for student disaffection. The four most frequently mentioned engagement factors were (a) problem-centric learning, (b) active learning supported by timely feedback, (c) course resources that cater to participants’ learnings need or preferences, and (d) instructor attributes such as enthusiasm or humor. The two most commonly reported student disaffections across the 10 MOOCs were due to forum- and peer-related issues. Conclusion This article ends with five main implications that could provide practical guidelines to other instructors of large online courses. The findings may also offer tips for instructors of traditional e-learning courses. Although we cannot generalize the findings of this study to traditional e-learning courses, it is possible that at the very least, the information presented here may suggest probable solutions for traditional e-learning courses that might otherwise be overlooked.