There are also numerous Youtube and other online resources on the Cynefin framework (ensure you select good quality reputable sources).
BUS9040M: Decision Analysis for Managers
Assignment 2: From Data to Decisions: using data analytics to support decision-making Introduction
Data-driven decision-making (DDDM) is the process of using data to inform decision-making. Data analytics is at the heart of DDDM. As discussed in the seminars, data analytics refers to the process and practice of analysing data to answer questions or to extract meaningful insights or information that an organization can use to inform its strategy and, ultimately, reach its objectives. In the context of the DIK pyramid (Wallace, 2007) (see seminar 1 and 2), managers can then use this information in combination with their experience and judgement to create knowledge and ultimately improve their decision-making.
In this assignment, your task is to independently apply data analytics techniques that you learn in the seminars to extract meaningful insights or information from data for decision making. In the context of the DIK pyramid, you may then draw out some information/knowledge from the data with a view to answering specific questions. Follow the steps below.
Step 1: Questions: Develop some 3 to 5 (research) questions that you seek to answer from data. These should be questions that may be of interest to managers or to an organisation seeking to extract useful insights, information or knowledge from data (see introduction above).
Step 2: Data: Find some data relevant to the questions you seek to address in step 1 (see guidance on potential data sources in the Appendix below) and download it onto an Excel spreadsheet.
You may have to “clean” the data before analysis (step 3). Data cleaning includes checking that the data is usable, ensuring that missing values are not too many, removing irrelevant data/variables, transforming existing variables or creating new variables, renaming variables, etc. You must upload your (clean) Excel data onto blackboard (see Appendix). [Needless to say, DO NOT use data provided in the seminars; find your own data! Some guidance on potential data sources provided in the appendix]
Step 3: Data analysis and report writing: Analyse the data and write a report. Your report should interpret/discuss the results with a view to answering the questions that you set out in step 1. (your report should be approximately 1,500 words, excluding Tables, Figures/charts and references)
Your analysis and report must include the following:
(a) A selection of descriptive analytics (numerical measures) appropriate for your data, questions or information required from the data – for guidance, see seminar 1.
(b)A selection of descriptive analytics (data visualization in the form of charts/graphs) appropriate for your data, questions or information required from the data – for guidance, see seminar 2.
(c) Predictive analytics (regression analysis): Based on your learning in seminar 3, use regression techniques to investigate a potential cause-and-effect relationship between your variables i.e., the relationship between your identified dependent variable and independent variables. For guidance, see box below and seminar 3.
Apart from (a) to (c) above, you may include any other data analytics techniques that you learn in the seminar series that are appropriate for your data, questions or information required from the data (if the word limits allow).
Note: You must name the source of your data in your report. Most data are publicly available. However, if it is sensitive data or your source wishes to remain anonymous (e.g., privileged data from a company), you may anonymise the source.
Some guidance:
(i) Before you conduct regression analysis, generate descriptive statistics of your dependent (response) variable and your independent variable(s) and summarise the results in a table following the example in Seminar 3 Worksheet Table 2 (report the mean, median and standard deviation figures to 2 or 3 decimal places). Do not simply copy and paste (unedited) Excel results output – construct your own table and populate it with only the mean, median and standard deviation values from Excel, following the example in Seminar 3 Worksheet Table 2. Briefly comment on the results in your report.
(ii) In addition, before conducting regression analysis, it is also useful to investigate how your variables are correlated by conducting a correlation analysis – this should be reported in a correlation matrix/table following the example in Seminar 3 Worksheet Table 3 (report correlation figures to 2 or 3 decimal places).
(iii) After you report descriptive statistics and correlation analysis results, conduct regression analysis to investigate the relationship between your dependent (response) variable and your independent variable(s). Summarise the results in a Table following the example in Seminar 3 Worksheet Table 4 (report the coefficient and standard error figures in 2 decimal places, while the p-values should be reported in 3 decimal places). Again, do not simply copy and paste (unedited) Excel results output – construct your own table and populate it with only the coefficient, standard error and p-values from Excel, following the example in Seminar 3 Worksheet Table 4.
(iv) Ensure you clearly report the regression equation for predicting your dependent variable and interpret the results, following the structure/guidance in Seminar 3 Worksheet. Then use your estimated regression equation to predict the value of the dependent variable based on the MEAN values of the independent variables that you reported in (i) above.
(v) Use information from the regression results to provide answers(s) to the relevant question(s) you set out in step 1 and draw some conclusions or insights from the data that may be useful for decision making.
(vi)Note that this part of your report (regression analysis) carries a relatively higher weight; therefore, if you do not include regression analysis, your report will be assessed as incomplete.
More guidance on writing the report
Your report should be structured around your answers to the questions you set out in step 1. That is, the questions and your answers should form sub-sections of your report. Essentially, the report is a synthesis of the results of your analysis (step 3). It should be well presented and structured and guided by the principles of good academic writing. Importantly, the report should be coherent and logically flow through the steps above i.e., from questions (step 1) through to analysis of data to effectively synthesising the results with a view to answering the questions, and finally drawing some conclusions or implications from the results for decision making (step 3) in the context of the DIK pyramid. Again, structure your report around your answers to the questions you set out in step 1.
Appendix: Guidance on data and sources
There are many sources of raw data that may be used to address your research questions. Ensure that the data (and your research questions) are relevant to your field of study (i.e., your Masters degree subject area) or to the field of business/management and that you appropriately acknowledge your data sources. Below are some sources of data. This is not exhaustive list;
(1) The “UK Data Service” is one of the University library’s databases for data from a wide range of sectors. Researchers may access open data collections without the need to register or login: https://ukdataservice.ac.uk/find-data/access-conditions/open-access/
Go to university library website; under Find, click Databases, then find UK Data Service. Alternatively, this link will take you directly to all Databases (A-Z Databases (lincoln.ac.uk). Then find the relevant database (i.e., UK Data Service) and login using your usual University login details and you will be ready to search your data.
(2) Publicly available data e.g.,
World Bank: https://data.worldbank.org/,
Eurostat: https://ec.europa.eu/eurostat/data/database FAO: https://www.fao.org/faostat/en/#home
Office of National Statistics (ONS) for the UK or equivalent for other countries.
(3) If you are interested in financial data, particularly stock market data, you may access historical stock market data for any of the world’s major indices from:
https://uk.finance.yahoo.com/ OR
https://uk.investing.com/indices/
(4) Kaggle, a subsidiary of Google LLC, is an online community of data scientists. Among many things, Kaggle allows users to find and publish data sets. https://www.kaggle.com/
(5) Of course, you may use your own data sources e.g., if you have access to relevant data from your own country, an organisation that you previously worked for or one you are familiar with. However, ensure the data is relevant and of good quality data.
Finally, ensure you have a reasonable sample size of (raw) data (a sample size of least 30 to around 100 observations or lines of data). Of course you may use larger datasets if available – the more the better!. Needless to say, the data should be relatively recent (perhaps during the last 20 years), unless you are using historical time series data from the most recent going back in time.
NOTE:You MUST separately upload onto blackboard your Excel spreadsheet containing your RAW data as part of your submission. There will be a separate submission point specifically for the data used in this assignment (separate from the main assignment submission). If you DO NOT upload your data, your assignment will be assessed as incomplete.
