Understanding Data and their Environment

DATA71011 Understanding Data and their Environment 15 Semester 1 Postgraduate Taught Level 7 Pradyumn Shukla Unit coordinator Mark Elliot Unit coordinator Stian Soiland-Reyes Unit coordinator Nuno Pinto Unit coordinator Alliance Manchester Business School Social Statistics Planning, Property and Environmental Management Department of Computer Science 1 FHEQ level (Framework for Higher Education Qualifications) ' Masters/Integrated Masters P4 ' European Credit Transfer & Accumulation System Rating : 7.5 This module is a combination of technical and non-technical topics all related to critical externalities to the data analytics process. The course covers a suite of topics rated to the representation and processing and pre-processing of data: metadata, paradata ,data provenance: understanding data quality and the impact on inference; cleaning data; edit and imputation models, the basics of data linkage/integration and data visualisation.  This module is a combination of technical and non-technical topics all related to critical externalities to the data analytics process. The primary aim of the module is to demonstrate that data science cannot be carried out in a vacuum that a whole range of extrinsic considerations affect our ability to carry out the research that we wish to carry out.   However, appropriate management of these externalities can lead to higher quality as well more responsible research. The course will have 4 components:  1. Information about Data: metadata, paradata and data provenance. Provenance; Issues about data quality and the impact on inference; accessing and finding data.  2. Pre-Processing:  Understanding data quality and divergence and the impact on inference; Cleaning data; Editing and imputation models.  3. Combining and enhancing data: Basics of data linkage/integration.  4. Data Visualisation. All weeks have a one hour lectured followed by a two practical session with guided computer based practical exercises.  Week 8 is a team-based exercise where the teams are set a challenge involving bring multiple datasets together to build a model. See above the provisional timetable. You will stay in your groups for the first part of the independent study for the assessed work. As a group you will produce a data processing and analysis plan and give a group presentation of that plan in week 12 and receive feedback on the plan. You will then be given one week during which you can continue to discuss and compare notes. Thereafter you must work individually to produce your assessed report. You are free to vary from the group analysis plan in your individual reports.  The unit aims to:  • Develop an awareness of the issues around the use of data in research.  • Develop fundamental skills in data pre-processing.   Students should be able to:  • Demonstrate a basic understanding of metadata, paradata and data provenance  • Be able to prepare a dataset for analysis  • Make informed decisions about linkage/integration of data and carry out a basic data linkage.  • Be able to produce basic data visualisations.  Lectures will introduce specific ideas in relation to data management, the ethics and disclosure of data and linkage in relation to research. Interactive exercises will involve a mixture of solo and group work. Laptop based practicals will allow the students to apply those ideas and to manage data and be able to make informed decisions about linkage/integration of data and to apply anonymisation processes to data.  0 Other 100% Group provenance exercise (600 words and code) 20%Online test on information about data and reproducibility (1 hour) 20%Group presentation of analysis plan (video) (5 minutes) 10%Pre-processing and analysis report (1,500 words) 50%  Christen, P. (2012). Data matching: concepts and techniques for record linkage,   entity resolution, and duplicate detection. Springer Science & Business Media   García S., Luengo, J., & Herrera F. (2015). Data preprocessing in data mining.   Springer  Moreau, L., & Groth, P. (2013) Provenance: An Introduction to PROV. Available at https://tinyurl.com/PROV-BOOK [accessed 25/9/2019]    0 0 0