Can prospective usability evaluation predict data errors?

Increasing amounts of clinical research data are collected by manual data entry into electronic source systems and directly from research subjects. For this manual entered source data, common methods of data cleaning such as post-entry identification and resolution of discrepancies and double data entry are not feasible. However data accuracy rates achieved without these mechanisms may be higher than desired for a particular research use. We evaluated a heuristic usability method for utility as a tool to independently and prospectively identify data collection form questions associated with data errors. The method evaluated had a promising sensitivity of 64% and a specificity of 67%. The method was used as described in the literature for usability with no further adaptations or specialization for predicting data errors. We conclude that usability evaluation methodology should be further investigated for use in data quality assurance.

Quality Data is Key to Improving Education

The Data Quality Campaign (DQC) has been focused since 2005 on advocating for states to build robust state longitudinal data systems (SLDS). While states have made great progress in their data infrastructure, and should continue to emphasize this work, t data systems alone will not improve outcomes. It is time for both DQC and states to focus on building capacity to use the information that these systems are producing at every level – from classrooms to state houses. To impact system performance and student achievement, the ingrained culture must be replaced with one that focuses on data use for continuous improvement. The effective use of data to inform decisions, provide transparency, improve the measurement of outcomes, and fuel continuous improvement will not come to fruition unless there is a system wide focus on building capacity around the collection, analysis, dissemination, and use of this data, including through research.

Using Provenance for Quality Assessment and Repair in Linked Open Data

As the number of data sources publishing their data on the Web of Data is growing, we are experiencing an immense growth of the Linked Open Data cloud. The lack of control on the published sources, which could be untrustworthy or unreliable, along with their dynamic nature that often invalidates links and causes conflicts or other discrepancies, could lead to poor quality data. In order to judge data quality, a number of quality indicators have been proposed, coupled with quality metrics that quantify the “quality level” of a dataset. In addition to the above, some approaches address how to improve the quality of the datasets through a repair process that focuses on how to correct invalidities caused by constraint violations by either removing or adding triples. In this paper we argue that provenance is a critical factor that should be taken into account during repairs to ensure that the most reliable data is kept. Based on this idea, we propose quality metrics that take into account provenance and evaluate their applicability as repair guidelines in a particular data fusion setting.

Use of phosphorous mapping and sea-level change data in assessing coastal activity zones and establishing site chronology : A case study from the Icelandic multi-period site of Vatnsfjörður

This paper forms part of Lukasz Mikolajczyk's PhD dissertation, which is supervised by Karen Milek

Towards a reverse engineering approach for guiding user in applying data mining

Comunicación presentada en las XVI Jornadas de Ingeniería del Software y Bases de Datos, JISBD 2011, A Coruña, 5-7 septiembre 2011.

Assessing the quality of risk factor survey data: lessons from the WHO MONICA Project

Background and purpose Survey data quality is a combination of the representativeness of the sample, the accuracy and precision of measurements, data processing and management with several subcomponents in each. The purpose of this paper is to show how, in the final risk factor surveys of the WHO MONICA Project, information on data quality were obtained, quantified, and used in the analysis. Methods and results In the WHO MONICA (Multinational MONItoring of trends and determinants in CArdiovascular disease) Project, the information about the data quality components was documented in retrospective quality assessment reports. On the basis of the documented information and the survey data, the quality of each data component was assessed and summarized using quality scores. The quality scores were used in sensitivity testing of the results both by excluding populations with low quality scores and by weighting the data by its quality scores. Conclusions Detailed documentation of all survey procedures with standardized protocols, training, and quality control are steps towards optimizing data quality. Quantifying data quality is a further step. Methods used in the WHO MONICA Project could be adopted to improve quality in other health surveys.

Visualisation of quality information for geospatial and remote sensing data:providing the GIS community with the decision support tools for geospatial dataset quality evaluation

The evaluation of geospatial data quality and trustworthiness presents a major challenge to geospatial data users when making a dataset selection decision. The research presented here therefore focused on defining and developing a GEO label – a decision support mechanism to assist data users in efficient and effective geospatial dataset selection on the basis of quality, trustworthiness and fitness for use. This thesis thus presents six phases of research and development conducted to: (a) identify the informational aspects upon which users rely when assessing geospatial dataset quality and trustworthiness; (2) elicit initial user views on the GEO label role in supporting dataset comparison and selection; (3) evaluate prototype label visualisations; (4) develop a Web service to support GEO label generation; (5) develop a prototype GEO label-based dataset discovery and intercomparison decision support tool; and (6) evaluate the prototype tool in a controlled human-subject study. The results of the studies revealed, and subsequently confirmed, eight geospatial data informational aspects that were considered important by users when evaluating geospatial dataset quality and trustworthiness, namely: producer information, producer comments, lineage information, compliance with standards, quantitative quality information, user feedback, expert reviews, and citations information. Following an iterative user-centred design (UCD) approach, it was established that the GEO label should visually summarise availability and allow interrogation of these key informational aspects. A Web service was developed to support generation of dynamic GEO label representations and integrated into a number of real-world GIS applications. The service was also utilised in the development of the GEO LINC tool – a GEO label-based dataset discovery and intercomparison decision support tool. The results of the final evaluation study indicated that (a) the GEO label effectively communicates the availability of dataset quality and trustworthiness information and (b) GEO LINC successfully facilitates ‘at a glance’ dataset intercomparison and fitness for purpose-based dataset selection.

A quality driven adaptation scheme for DASH streaming

In this work we present a quality driven approach to DASH (Dynamic Adaptive Streaming over HTTP) for segment selection in varying network conditions. Current adaption algorithms focus largely on regulating data rates using network layer parameters by selecting the level of quality on offer that can eliminate buffer underrun without considering picture fidelity. In reality, viewers may accept a level of buffer underrun in order to achieve an improved level of picture fidelity. In this case, the conventional DASH algorithms can cause extreme degradation of the picture fidelity when attempting to eliminate buffer underrun with scarce bandwidth availability. Our work is concerned with a quality-aware rate adaption scheme that maximizes the client's quality of experience in terms of both continuity and fidelity (picture quality). Results show that the scheme proposed can maintain a high level of quality for streaming services, especially at low packet loss rates. It is also shown that by eliminating buffer underrun completely, the PSNR that reflects the picture quality of the video is greatly reduced. Our scheme offers the offset between continuity-based quality and resolution-based quality, which can be used to set threshold values for the level of quality desired by clients with different quality requirements. © 2013 IEEE.

Content-aware compression for big textual data analysis

A substantial amount of information on the Internet is present in the form of text. The value of this semi-structured and unstructured data has been widely acknowledged, with consequent scientific and commercial exploitation. The ever-increasing data production, however, pushes data analytic platforms to their limit. This thesis proposes techniques for more efficient textual big data analysis suitable for the Hadoop analytic platform. This research explores the direct processing of compressed textual data. The focus is on developing novel compression methods with a number of desirable properties to support text-based big data analysis in distributed environments. The novel contributions of this work include the following. Firstly, a Content-aware Partial Compression (CaPC) scheme is developed. CaPC makes a distinction between informational and functional content in which only the informational content is compressed. Thus, the compressed data is made transparent to existing software libraries which often rely on functional content to work. Secondly, a context-free bit-oriented compression scheme (Approximated Huffman Compression) based on the Huffman algorithm is developed. This uses a hybrid data structure that allows pattern searching in compressed data in linear time. Thirdly, several modern compression schemes have been extended so that the compressed data can be safely split with respect to logical data records in distributed file systems. Furthermore, an innovative two layer compression architecture is used, in which each compression layer is appropriate for the corresponding stage of data processing. Peripheral libraries are developed that seamlessly link the proposed compression schemes to existing analytic platforms and computational frameworks, and also make the use of the compressed data transparent to developers. The compression schemes have been evaluated for a number of standard MapReduce analysis tasks using a collection of real-world datasets. In comparison with existing solutions, they have shown substantial improvement in performance and significant reduction in system resource requirements.

Ensemble Stream Model for Data-Cleaning in Sensor Networks

Ensemble Stream Modeling and Data-cleaning are sensor information processing systems have different training and testing methods by which their goals are cross-validated. This research examines a mechanism, which seeks to extract novel patterns by generating ensembles from data. The main goal of label-less stream processing is to process the sensed events to eliminate the noises that are uncorrelated, and choose the most likely model without over fitting thus obtaining higher model confidence. Higher quality streams can be realized by combining many short streams into an ensemble which has the desired quality. The framework for the investigation is an existing data mining tool. First, to accommodate feature extraction such as a bush or natural forest-fire event we make an assumption of the burnt area (BA*), sensed ground truth as our target variable obtained from logs. Even though this is an obvious model choice the results are disappointing. The reasons for this are two: One, the histogram of fire activity is highly skewed. Two, the measured sensor parameters are highly correlated. Since using non descriptive features does not yield good results, we resort to temporal features. By doing so we carefully eliminate the averaging effects; the resulting histogram is more satisfactory and conceptual knowledge is learned from sensor streams. Second is the process of feature induction by cross-validating attributes with single or multi-target variables to minimize training error. We use F-measure score, which combines precision and accuracy to determine the false alarm rate of fire events. The multi-target data-cleaning trees use information purity of the target leaf-nodes to learn higher order features. A sensitive variance measure such as f-test is performed during each node’s split to select the best attribute. Ensemble stream model approach proved to improve when using complicated features with a simpler tree classifier. The ensemble framework for data-cleaning and the enhancements to quantify quality of fitness (30% spatial, 10% temporal, and 90% mobility reduction) of sensor led to the formation of streams for sensor-enabled applications. Which further motivates the novelty of stream quality labeling and its importance in solving vast amounts of real-time mobile streams generated today.

Data analysis of collapse mechanisms of a 3D printed groin vault in shaking table testing

The aim of this novel experimental study is to investigate the behaviour of a 2m x 2m model of a masonry groin vault, which is built by the assembly of blocks made of a 3D-printed plastic skin filled with mortar. The choice of the groin vault is due to the large presence of this vulnerable roofing system in the historical heritage. Experimental tests on the shaking table are carried out to explore the vault response on two support boundary conditions, involving four lateral confinement modes. The data processing of markers displacement has allowed to examine the collapse mechanisms of the vault, based on the arches deformed shapes. There then follows a numerical evaluation, to provide the orders of magnitude of the displacements associated to the previous mechanisms. Given that these displacements are related to the arches shortening and elongation, the last objective is the definition of a critical elongation between two diagonal bricks and consequently of a diagonal portion. This study aims to continue the previous work and to take another step forward in the research of ground motion effects on masonry structures.

Brazilian flavonoid database: Application of quality evaluation system

Much information on flavonoid content of Brazilian foods has already been obtained; however, this information is spread in scientific publications and non-published data. The objectives of this work were to compile and evaluate the quality of national flavonoid data according to the United States Department of Agriculture`s Data Quality Evaluation System (USDA-DQES) with few modifications, for future dissemination in the TBCA-USP (Brazilian Food Composition Database). For the compilation, the most abundant compounds in the flavonoid subclasses were considered (flavonols, flavones, isoflavones, flavanones, flavan-3-ols, and anthocyanidins) and the analysis of the compounds by HPLC was adopted as criteria for data inclusion. The evaluation system considers five categories, and the maximum score assigned to each category is 20. For each data, a confidence code (CC) was attributed (A, B, C and D), indicating the quality and reliability of the information. Flavonoid data (773) present in 197 Brazilian foods were evaluated. The CC ""C"" (as average) was attributed to 99% of the data and ""B"" (above average) to 1%. The main categories assigned low average scores were: number of samples; sampling plan and analytical quality control (average scores 2, 5 and 4, respectively). The analytical method category received an average score of 9. The category assigned the highest score was the sample handling (20 average). These results show that researchers need to be conscious about the importance of the number and plan of evaluated samples and the complete description and documentation of all the processes of methodology execution and analytical quality control. (C) 2010 Elsevier Inc. All rights reserved.

Brazilian Network of Food Data Systems and LATINFOODS Regional Technical Compilation Committee: Food composition activities (2006-2009)

The Brazilian Network of Food Data Systems (BRASILFOODS) has been keeping the Brazilian Food Composition Database-USP (TBCA-USP) (http://www.fcf.usp.br/tabela) since 1998. Besides the constant compilation, analysis and update work in the database, the network tries to innovate through the introduction of food information that may contribute to decrease the risk for non-transmissible chronic diseases, such as the profile of carbohydrates and flavonoids in foods. In 2008, data on carbohydrates, individually analyzed, of 112 foods, and 41 data related to the glycemic response produced by foods widely consumed in the country were included in the TBCA-USP. Data (773) about the different flavonoid subclasses of 197 Brazilian foods were compiled and the quality of each data was evaluated according to the USDAs data quality evaluation system. In 2007, BRASILFOODS/USP and INFOODS/FAO organized the 7th International Food Data Conference ""Food Composition and Biodiversity"". This conference was a unique opportunity for interaction between renowned researchers and participants from several countries and it allowed the discussion of aspects that may improve the food composition area. During the period, the LATINFOODS Regional Technical Compilation Committee and BRASILFOODS disseminated to Latin America the Form and Manual for Data Compilation, version 2009, ministered a Food Composition Data Compilation course and developed many activities related to data production and compilation. (C) 2010 Elsevier Inc. All rights reserved.

The effects of normalisation on the ability of business end-users to detect data anomalies: An experimental evaluation

With the proliferation of relational database programs for PC's and other platforms, many business end-users are creating, maintaining, and querying their own databases. More importantly, business end-users use the output of these queries as the basis for operational, tactical, and strategic decisions. Inaccurate data reduce the expected quality of these decisions. Implementing various input validation controls, including higher levels of normalisation, can reduce the number of data anomalies entering the databases. Even in well-maintained databases, however, data anomalies will still accumulate. To improve the quality of data, databases can be queried periodically to locate and correct anomalies. This paper reports the results of two experiments that investigated the effects of different data structures on business end-users' abilities to detect data anomalies in a relational database. The results demonstrate that both unnormalised and higher levels of normalisation lower the effectiveness and efficiency of queries relative to the first normal form. First normal form databases appear to provide the most effective and efficient data structure for business end-users formulating queries to detect data anomalies.

SmartClean: an incremental data cleaning tool

This paper presents the SmartClean tool. The purpose of this tool is to detect and correct the data quality problems (DQPs). Compared with existing tools, SmartClean has the following main advantage: the user does not need to specify the execution sequence of the data cleaning operations. For that, an execution sequence was developed. The problems are manipulated (i.e., detected and corrected) following that sequence. The sequence also supports the incremental execution of the operations. In this paper, the underlying architecture of the tool is presented and its components are described in detail. The tool's validity and, consequently, of the architecture is demonstrated through the presentation of a case study. Although SmartClean has cleaning capabilities in all other levels, in this paper are only described those related with the attribute value level.