联系我们： 手动添加方式: 微信>添加朋友>企业微信联系人>13262280223 或者 QQ: 1483266981SOES6016 Assessment 1, 2022-23: Short Essay

Fundamentals of Data Science
Assessed Project Specification Preview
Project description
A technical report summarising your project should be submitted as a pdf document through
the Exeter Learning Environment (ELE) by the assessment deadline of 12 noon on 12th Jan_xfffe_uary. Note that late submissions will be penalised unless an extension is granted via the
standard mitigation process.
The first page of your report must include the statement “I have familiarised myself with
the academic misconduct and plagiarism guidelines in the Academic Honesty and Plagiarism
module and the MTHM601 (Fundamentals of Data Science) ELE site’s Assessment Information
tab. This report constitutes my own work, and I have explicitly referenced and acknowledged
those parts that draw on the literature, online sources, and the support of others. This includes
acknowledgment of any use of Artificial Intelligence tools such as ChatGPT.”
Your project will be on a topic chosen by you. It should involve analysis of a dataset to answer a specific
question or set of questions. You should use your dataset to demonstrate your mastery of the broad topics
we have covered in this course. In addition, you should demonstrate your ability to go beyond what you have
been taught and conduct some self-directed learning, e.g., use some R package not covered in the course,
or extend some of the ideas covered, or demonstrate a deeper understanding of a topic through your own
exploration.
You should:
1. Find some data. You can either collect it yourself or you can source it online. As a general rule,
the data you choose should have more than 100 observations (rows) and a mixture of variable types
(numeric, categorical, ordinal, and so forth). Feel free to merge existing datasets to create a new
dataset. Data is available in many places, and the details of some recommended sites are provided at
the end of this document.
2. You should formulate a question of interest (e.g., do countries with higher GDP also have higher
carbon dioxide emissions where in the UK is cycle commuting most prevalent ) that can be answered
with the data you have, and provide the most comprehensive answer to that question that you can
using the ideas covered in the course. Note that you can either find the data first and come up with the
research question later, or the reverse, or refine your research question as you go along. But you should
be sure to explain your methodology and be alert to good and bad practice in regards to p-hacking,
pre-specification and performing multiple comparisons.
3. Write a report summarising your findings. Your answer should be supported by graphics and
tables, as well as formal analysis. You should discuss your findings and interpret your results, including
the limitations of what you have found.
4. Put your code in an appendix. You should include an appendix containing the annotated code that
you used to answer the question, or submit your code over the OneDrive by placing it in a directory and
sharing the directory with Ed and Tim (E.H.Branford@exeter.ac.uk and T.H.Hughes@exeter.ac.uk).
This can be R code or the contents of a .Rmd file. In particular, it is important that your work
1
be reproducible (e.g., you share the details of your raw data and executable code for wrangling this,
producing your visualisations, and conducting your analysis). The assessment will be based on the
report itself, so if you wish to draw attention to a specific feature of your code then please provide
code extracts within the report itself, or references to the code location on OneDrive, with a clear
explanation of what the code does and how to execute it.
Your report should be no longer than 10 pages including figures and tables, but excluding your
appendix. If you have referred to the literature or other sources when producing your project then you must
provide appropriate references, and make use of paraphrasing or block quotations when summarising other
peoples’ work. You may use any software you like to create the report (including Word), but RMarkdown is
recommended. Whichever software you use, it is necessary to submit the report through ELE in pdf format
as a single file, with any supplementary material referenced in the report itself and submitted over OneDrive
prior to the assessment deadline.
Here is a suggested structure with some suggested page lengths:
1. Introduction: here you should introduce the question you are interested in answering. You could give
some background on the problem/question and list which topic areas you will showcase. ～ 1 page.
2. Objectives/Methodology: here you should set out what you plan to do. You should have a research
question you are hoping to answer and some objectives or a methodology to follow to help you get
there. ～ 1 page.
3. Data: here you could describe the data that you will use, where you got it from and briefly describe
what wrangling you may have had to do to format it correctly. ～ 1 page.
4. Results: here you should present the results of your analysis. Figures and tables should be presented
as you go along. You should include discussion as you go along. ～ 5 pages.
5. Limitations: here you can highlight any limitations of your work, what likely impact those limitations
have and how you might address those limitations in a perfect world. ～ 1 page.
6. Conclusion: here you should summarise your findings succinctly. ～ 1 page.
Some general hints and tips:
You should provide your student number at the top of the page and also indicate which topics you are
demonstrating.
You should demonstrate your understanding of the concepts and techniques covered in the course and
follow the best practice advice offered in the sessions and the pre-recorded lecture videos.
If you include a figure/table, make sure it is discussed. If it isn’t, that indicates that it’s not needed.
It is a good idea to aim to have no more than one figure or table per page.
Write clearly and distinctly using short sentences.
Pay attention to how the final report looks. Make sure figure labels and legends are readable.
Make sure there is a logical structure to the report so that a reader understands what you are doing
and why you are doing it at every stage.
Your report should be a readable document, so code should be in the appendix, not in the report,
unless including an extract of code to highlight an important point.
2
Get started early. Start writing as soon as you can so that problems with compiling the document or
downloading R packages are sorted quickly.
Guidance on how to reference and avoid plagiarism is available on ELE, in particular the 30-60 minute
module called “Academic Honesty and Plagiarism”: https://ele.exeter.ac.uk/course/view.php id=1959.
Recommended datasets
There are a number of excellent websites sharing open source data that you may wish to explore in your
projects and for general interest. The following is a list of some of these websites together with a brief
description. This list is by no means exhaustive, and we welcome further suggestions and look forward to
see what other data you source or collect over the programme.
Kaggle https://www.kaggle.com/datasets. This is a repository of community published datasets and
code. It offers access to free graphics processing units (performance computing) which can be accessed
via a user interface that allows you to explore the datasets in either Python or R and write your own
literate programming worksheets using markdown.
UN Data http://data.un.org/Explorer.aspx. This is a data service compiled by the United Nations
statistics division and other international agencies. Application areas include agriculture, communica tion, development assistance, education, energy, environment, finance, gender, health, labour market,
manufacturing, national accounts, population and migration, science and technology, tourism, trans port and trade. The data is available to download in csv format alongside other formats. It is
recommended to click on “Topics” on the explorer page to search for data by application area.
Gapminder https://www.gapminder.org/. This is a foundation with the premise of fighting miscon ceptions and promoting a fact-based worldview. Time series data are available under the categories of
communication, economy, education, energy, environment, health, infrastructure, population, society
and work. They also host data on World Development Indicators copied from the World Bank. Be sure
to follow the guidance on attributing the original data source and the Gapminder foundation itself. The
data itself can be accessed at the page https://www.gapminder.org/data/, and details on how missing
data have been handled is available at the page https://www.gapminder.org/data/documentation/.
The Global Health Observatory of the World Health Organisation https://apps.who.int/gho/data/
node.home. This contains data on health status indicators to monitor progress towards the health
objectives of the Sustainable Development Goals. Data is available to download in csv format, alongside
other formats.
Our World in Data https://ourworldindata.org/. This contains charts of progress against global devel opment indicators, such as the Sustainable Development Goals. A valuable source of examples of good
data visualisation practice. Links are provided to other websites that host the original data sources.
European Environment Agency https://www.eea.europa.eu/en/analysis. This is an agency of the Eu ropean Union that aims to support sustainable development by helping to achieve significant and
measurable improvement in Europe’s environment, through the provision of timely, targeted, relevant
and reliable information to policymaking agents and the public.
Eurostat https://ec.europa.eu/eurostat/data/database. The statistical office of the European Union,
co-ordinating statistical activities across the European Commission. Maintain an inventory of statisti cal data collections of conventional indicators. And have a number of visualisation tools and educational
resources.
3
Google Datasets https://datasetsearch.research.google.com/. This is a search engine for datasets across
a wide range of online repositories. Its mission is also to foster a data sharing ecosystem that will
encourage data publishers to follow best practices for data storage and publication, and to allow
scientists to show the impact of their work through citation of datasets that they have produced
UK Data Service https://www.ukdataservice.ac.uk/. A repository of UK census and related data.
Note that it is necessary to request access to many of the UK Data Service datasets, so please plan
accordingly.
Office for National Statistics https://www.ons.gov.uk/. UK’s recognised national statistical institute.
Collect and publish statistics related to economy, population and society in the UK, and conduct the
census in England and Wales every 10 years.
The National Snow and Ice Data Center https://nsidc.org/data. An open source database of glacier
and sea ice extent and other environmental data relating to the Arctic, Antarctic and Greenland.
4