EC994: Applications of Data Science

Nathan Canen
Module Leader

15 CATS - Department of Economics

Summer Module

Spring Module

Introduction

EC994-15 Applications of Data Science

Principal Aims

Big data is transforming almost every aspect of science and the humanities, driven by the emergence of a data society. This is a society in which increasingly comprehensive aspects of human behaviour and the economy are recorded as data. Employers are recognizing the need for a skilled workforce that can extract value from data, giving rise to the new job description of a data scientist. This course aims to provide economists and social scientist with a solid basis to overcome the deep technical deficit that has been identified among social scientists in the methodologies and practical tools of data science (Rebekah Luff, Rose Wiles and Patrick Sturgis, “Consultation on Methodological Research Needs in UK Social Science”, National Centre for Research Methods, March 2015.)The aim of this module is provide students with a thorough understanding of the most common statistical methods related to high-dimensional data and machine learning techniques, with a particular focus to applications on economic and social data. The course will cover both the theory underpinning these methods and will also feature an intensive applied computing component.

Principal Learning Outcomes

Subject Knowledge and Understanding:...demonstrate awareness and understanding of key methods available for statistical learning and dimensionality reduction (Lasso, SVM, Networks, Bagging, Clustering). The teaching and learning methods that enable students to achieve this learning outcome are: Lectures, seminars with applied modules, independent study. The summative assessment methods that measure the achievement of this learning outcome are: Examination.

Subject Knowledge and Understanding:...demonstrate an understanding of how these methods may be used to in different contexts. The teaching and learning methods that enable students to achieve this learning outcome are: Lectures, seminars with applied modules, independent study. The summative assessment methods that measure the achievement of this learning outcome are: Examination.

Subject-specific skills/Professional Skills:...gain an understanding for and an ability to differentiate the appropriateness of different statistical methods. The teaching and learning methods that enable students to achieve this learning outcome are: Lectures, seminars with applied modules, independent study. The summative assessment methods that measure the achievement of this learning outcome are: Examination.

Subject-specific skills/Professional Skills An ability to apply data science methods to every day challenges. The teaching and learning methods that enable students to achieve this learning outcome are: Lectures, seminars with applied modules, independent study. The summative assessment methods that measure the achievement of this learning outcome are: Examination.

Syllabus

The syllabus may cover, but is not limited to, the following areas:

• Data Science Use cases (e.g. in academia, business, public sector)

• Linear Methods

• Naïve Bayes

• General Linear models

• Model selection

• Bootstrapping

• Random trees, forests

• Dimensionality reduction (Principal Component, Clustering)

• Supervised learning methods

• Unsupervised learning

• Applications using statistical packages (such as R or others)

Context

Optional Module: L1P6 - Year 1, L1P7 - Year 1, G300 - Year 3, G300 - Year 4, G1PF - Year 1, C8P7 - Year 1, C803 - Year 1
Pre or Co-requisites: Probability and statistics, including linear regression/OLS. Basic maths (Algebra, Analysis). Programming skills are helpful but not a prerequisite.

Assessment

Assessment Method: Centrally-timetabled examination (On-campus) (100%)
Coursework Details: Centrally-timetabled examination (On-campus) (100%)
Exam Timing: May

Exam Rubric

Time Allowed: 2 Hours

Read all instructions carefully - and read through the entire paper at least once before you start entering your answers.

There are TWO sections in this paper. Answer ALL FOUR questions in Section A (10 marks each) and TWO questions in Section B (30 marks each).

Use a separate booklet for each Section.

• Use a PINK booklet for Section A questions.

• Use a SEPARATE PINK booklet for Section B questions.

You must write the number(s) of the question(s) you have answered on the front cover of each booklet. Make sure the numbers are clearly visible and correspond to the questions you completed inside that booklet.

Do not submit answers to more than the required number of questions. If you do, only the first answers (in the order they appear) will be marked, up to the required number for each section.

Approved scientific (non-graphical) pocket calculators are allowed.

Previous exam papers can be found in the University’s past papers archive. Please note that previous exam papers may not have operated under the same exam rubric or assessment weightings as those for the current academic year. The content of past papers may also be different.

Nathan Canen