Bike sharing systems (BSSs) are becoming an important part of urban mobility in many cities given that they are sustainable and environmentally friendly. BSS operators spend great efforts to ensure bike and dock availability at each station. Measuring the quality of service (QoS) of each station and/or the entire system is critical for efficient system operations. The traditionally-known QoS measure reported in the literature is based on the proportion of problematic stations, which are defined as those with no bikes or docks available to users. This measure neither exposes the spatial dependencies between stations nor does it discriminate between stations in the BSS. Hence, we propose a novel QoS measure, namely the Optimal Occupancy, in which: 1) the temporal variations in arrival and pick up rates at individual stations are considered; 2) the discriminative property of the Optimal Occupancy is demonstrated using Analysis of Variance (ANOVA) procedures; and 3) geo-statistics, which have not been used before, are applied to explore the spatial Optimal Occupancy variations and model variograms for spatial prediction. This study uses an anonymized bike trip dataset from 34 stations in downtown San Francisco to compare the traditionally-known QoS measure and the proposed Optimal Occupancy measure. Results reveal that the Optimal Occupancy is beneficial, outperforms the traditionally-known QoS measure, and produces a better prediction of the QoS at nearby locations. In addition, the Optimal Occupancy can be used to predict candidate locations for the introduction of new stations in an existing BSS.

Interactive machine learning (ML) systems are difficult to design because of the “Two Black Boxes” problem that exists at the interface between human and machine. Many algorithms that are used in interactive ML systems are black boxes that are presented to users, while the human cognition represents a second black box that can be difficult for the algorithm to interpret. These black boxes create cognitive gaps between the user and the interactive ML model. In this paper, we identify several cognitive gaps that exist in a previously-developed interactive visual analytics (VA) system, Andromeda, but are also representative of common problems in other VA systems. Our goal with this work is to open both black boxes and bridge these cognitive gaps by making usability improvements to the original Andromeda system. These include designing new visual features to help people better understand how Andromeda processes and interacts with data, as well as improving the underlying algorithm so that the system can better implement the intent of the user during the data exploration process. We evaluate our designs through both qualitative and quantitative analysis, and the results confirm that the improved Andromeda system outperforms the original version in a series of high-dimensional data analysis tasks.

This paper develops a novel two-layer hierarchical classifier that increases the accuracy of traditional transportation mode classification algorithms. This paper also enhances classification accuracy by extracting new frequency domain features. Many researchers have obtained these features from global positioning system data; however, this data was excluded in this paper, as the system use might deplete the smartphone's battery and signals may be lost in some areas. Our proposed two-layer framework differs from previous classification attempts in three distinct ways: 1) the outputs of the two layers are combined using Bayes' rule to choose the transportation mode with the largest posterior probability; 2) the proposed framework combines the new extracted features with traditionally used time domain features to create a pool of features; and 3) a different subset of extracted features is used in each layer based on the classified modes. Several machine learning techniques were used, including k-nearest neighbor, classification and regression tree, support vector machine, random forest, and a heterogeneous framework of random forest and support vector machine. Results show that the classification accuracy of the proposed framework outperforms traditional approaches. Transforming the time domain features to the frequency domain also adds new features in a new space and provides more control on the loss of information. Consequently, combining the time domain and the frequency domain features in a large pool and then choosing the best subset results in higher accuracy than using either domain alone. The proposed two-layer classifier obtained a maximum classification accuracy of 97.02%.

Much research has been done regarding how to visualize and interact with observations and attributes of high-dimensional data for exploratory data analysis. From the analyst's perceptual and cognitive perspective, current visualization approaches typically treat the observations of the high-dimensional dataset very differently from the attributes. Often, the attributes are treated as inputs (e.g., sliders), and observations as outputs (e.g., projection plots), thus emphasizing investigation of the observations. However, there are many cases in which analysts wish to investigate both the observations and the attributes of the dataset, suggesting a symmetry between how analysts think about attributes and observations. To address this, we define SIRIUS (Symmetric Interactive Representations In a Unified System), a symmetric, dual projection technique to support exploratory data analysis of high-dimensional data. We provide an example implementation of SIRIUS and demonstrate how this symmetry affords additional insights.

Exploring high-dimensional data is challenging. Dimension reduction algorithms, such as weighted multidimensional scaling, support data exploration by projecting datasets to two dimensions for visualization. These projections can be explored through parametric interaction, tweaking underlying parameterizations, and observation-level interaction, directly interacting with the points within the projection. In this article, we present the results of a controlled usability study determining the differences, advantages, and drawbacks among parametric interaction, observation-level interaction, and their combination. The study assesses both interaction technique effects on domain-specific high-dimensional data analyses performed by non-experts of statistical algorithms. This study is performed using Andromeda, a tool that enables both parametric and observation-level interaction to provide in-depth data exploration. The results indicate that the two forms of interaction serve different, but complementary, purposes in gaining insight through steerable dimension reduction algorithms.

Dimension reduction algorithms and clustering algorithms are both frequently used techniques in visual analytics. Both families of algorithms assist analysts in performing related tasks regarding the similarity of observations and finding groups in datasets. Though initially used independently, recent works have incorporated algorithms from each family into the same visualization systems. However, these algorithmic combinations are often ad hoc or disconnected, working independently and in parallel rather than integrating some degree of interdependence. A number of design decisions must be addressed when employing dimension reduction and clustering algorithms concurrently in a visualization system, including the selection of each algorithm, the order in which they are processed, and how to present and interact with the resulting projection. This paper contributes an overview of combining dimension reduction and clustering into a visualization system, discussing the challenges inherent in developing a visualization system that makes use of both families of algorithms.

When high-dimensional data is visualized in a 2D plane by using parametric projection algorithms, users may wish to manipulate the layout of the data points to better reflect their domain knowledge or to explore alternative structures. However, few users are well-versed in the algorithms behind the visualizations, making parameter tweaking more of a guessing game than a series of decisive interactions. Translating user interactions into algorithmic input is a key component of Visual to Parametric Interaction (V2PI) [13]. Instead of adjusting parameters, users directly move data points on the screen, which then updates the underlying statistical model. However, we have found that some data points that are not moved by the user are just as important in the interactions as the data points that are moved. Users frequently move some data points with respect to some other 'unmoved' data points that they consider as spatially contextual. However, in current V2PI interactions, these points are not explicitly identified when directly manipulating the moved points. We design a richer set of interactions that makes this context more explicit, and a new algorithm and sophisticated weighting scheme that incorporates the importance of these unmoved data points into V2PI.