---
owner:
    oldhid: hid340
    firstname: Timothy
    lastname: Thompson
    hid: hid-sp18-705
    semester: Fall 2017
    course: Big Data Applications and Analytics (I523)
    online: Yes
    url: https://github.com/cloudmesh-community/hid-sp18-705
notebook:
    url: https://github.com/cloudmesh-community/hid-sp18-705/blob/master/notebook.md
paper1:
    abstract: >
        Research libraries and archives have played a longstanding
        role in information management and access. In the second half
        of the twentieth century, libraries were at the forefront of
        automation and networked access to information. Since the
        advent of the internet, however, they have failed to keep pace
        with technological advances and currently face serious
        challenges in serving the evolving needs of researchers, whose
        information-seeking strategies are now shaped by internet
        search engines and online social media applications. To remain
        relevant in the current information landscape, libraries and
        archives must implement new strategies for converting legacy
        metadata to new formats that can add value to the research
        process. Although the data and metadata produced by libraries
        and archives may not always qualify, prima facie, as big data,
        an awareness among information professionals of the tools,
        techniques, and affordances of big data can help make library
        services more relevant to researchers.
    author:
        - Timothy A. Thompson
    chapter: Media
    hid:
        - hid-sp18-705
    status: "100% Oct 25 17"
    title: Big Data Analytics for Research Libraries and Archives
    url: https://github.com/cloudmesh-community/hid-sp18-705/tree/master/paper1
paper2:
    review: "100% Nov 8 17"
    abstract: >
        Decentralized systems such as Bitcoin, the Interplanetary File
        System (IPFS), and Ethereum have been designed with the
        intention of reengineering the architecture of online
        information systems, minimizing exposure to centralized points
        of failure, and creating new social models for the exchange of
        data--which is posited as a valuable asset in and of itself.
        Can these kinds of systems also support big data analytics and
        processing? If so, what stands to be gained by taking a
        blockchain-based approach to big data? Efforts to integrate
        blockchains into big data pipelines must address the
        inevitable tradeoffs between security and scalability.
        BigchainDB is a new decentralized database framework that adds
        blockchain-based features, such as immutability and asset
        management, to traditional NoSQL distributed databases.
        Although it is still in the early stages of development,
        BigchainDB promises to make a significant contribution to the
        ways in which data is shared and managed at scale.
    author:
        - Timothy A. Thompson
    chapter: Technology
    hid:
        - hid-sp18-705
    status: "100% Nov 8 17"
    title: "BigchainDB: A Big Database for the Blockchain?"
    url: https://github.com/cloudmesh-community/hid-sp18-705/tree/master/paper2
project:
    review: Dec 4 2017
    author:
        - Timothy A. Thompson
    hid:
        - hid-sp18-705
    title: >
        New Approaches to Managing Metadata at Scale in Research
        Libraries
    abstract: >
        The analysis of big data often relies on distributed storage
        and computation; however, access to big data--and to the
        platforms capable of managing and processing it--continues to
        be largely centralized. Centralization is particularly evident
        in the case of the metadata produced, managed, and disseminated
        by academic and research libraries. Libraries typically create
        and share their catalog records by uploading them to a
        centrally managed database, which can then be searched by other
        libraries for records that can be copied and added to an
        institution's local catalog. This centralized approach, which
        operates on the basis of membership fees, has the advantage of
        scalability and availability, but it comes at the cost of a
        loss of autonomy. Although technical innovation is possible
        within the current paradigm, the growing maturity of peer-to-
        peer protocols and decentralized solutions points toward an
        alternative approach, one that would allow libraries to share
        their data directly without having to pay an expensive
        intermediary.
    url: https://github.com/cloudmesh-community/hid-sp18-705/tree/master/project-report
    type: project
    status: "100% Nov 29 18"
    chapter: Media
experiment:
    url: https://github.com/cloudmesh-community/hid-sp18-705/tree/master/experiments/perceptron
    title: Write Your Own Perceptron
    status: "100% Nov 26 17"