The key frameworks and concepts covered in modules 1-10 are particularly relevant for this assignment. Assignment 3 relates to the specific course learning objectives 1, 2, 3 and 4:
1. apply knowledge of people, markets, finances, technology and management in a global context of business intelligence practice (data warehousing and big data architecture, data mining process, data visualisation and performance management) and resulting organisational change and understand how these apply to the implementation of business intelligence in organisation systems and business processes
2. identify and solve complex organisational problems creatively and practically through the use of business intelligence and critically reflect on how evidence based decision making and sustainable business performance management can effectively address real-world problems
3. comprehend and address complex ethical dilemmas that arise from evidence based decision making and business performance management
4. communicate effectively in a clear and concise manner in written report style for senior management with the correct and appropriate acknowledgment of the main ideas presented and discussed.
Task 1
The goal of Task 1 is to predict the likelihood of rainfall for tomorrow (next day) based on today's weather conditions. In Task 1 of Assignment 3 you are required to use the data mining tool RapidMiner to analyse and report on the weatherAUS.csv data set provided for Assignment 3. You should review the data dictionary for weatherAUS.csv data set (see Table 1 below).
Table 1 Data dictionary for Australian Weather Data set variables
|
Variable Name
|
Data Type
|
Description
|
|
Date
|
Date
|
Date of weather observation
|
|
Location
|
Text
|
Common name of the location of the weather station.
|
|
MinTemp
|
Real
|
Minimum temperature in degrees Celsius.
|
|
MaxTemp
|
Real
|
Maximum temperature in degrees Celsius.
|
|
Rainfall
|
Real
|
Amount of rainfall recorded for the day in mm.
|
|
Evaporation
|
Real
|
So-called Class A pan evaporation (mm) in the 24 hours to 9am.
|
|
Sunshine
|
Real
|
Number of hours of bright sunshine in the day.
|
|
WindGustDir
|
Polynominal
|
Direction of the strongest wind gust in the 24 hours to midnight.
|
|
WindGustSpeed
|
Integer
|
Speed (km/h) of the strongest wind gust in the 24 hours to midnight.
|
|
WindDir9am
|
Polynominal
|
Direction of wind at 9am
|
|
WindDir3pm
|
Polynominal
|
Direction of wind at 3pm
|
|
WindSpeed9am
|
Integer
|
Wind speed (km/hr) averaged over 10 minutes prior to 9am.
|
|
WindSpeed3pm
|
Integer
|
Wind speed (km/hr) averaged over 10 minutes prior to 3pm.
|
|
Humidity9am
|
Integer
|
Relative humidity (percent) at 9am.
|
|
Humidity3pm
|
Integer
|
Relative humidity (percent) at 3pm.
|
|
Pressure9am
|
Real
|
Atmospheric pressure (hpa) reduced to mean sea level at 9am.
|
|
Pressure3pm
|
Real
|
Atmospheric pressure (hpa) reduced to mean sea level at 3pm.
|
|
Cloud9am
|
Integer
|
Fraction of sky obscured by cloud at 9am. This is measured in "oktas", which are a unit of eighths. It records how many eights of the sky are obscured by cloud. A 0 measure indicates completely clear sky whilst an 8 indicates that it is completely overcast.
|
|
Cloud3pm
|
Integer
|
Fraction of sky obscured by cloud (in "oktas": eighths) at 3pm. See Cload9am for a description of the values.
|
|
Temp9am
|
Real
|
Temperature (degrees C) at 9am.
|
|
Temp3pm
|
Real
|
Temperature (degrees C) at 3pm.
|
|
RainToday
|
Nominal
|
Integer: Yes if precipitation (mm) in the 24 hours to 9am exceeds 1mm, otherwise No.
|
|
RISK_MM
|
Real
|
Amount of rain. A kind of measure of the "risk".
|
|
RainTomorrow
|
Nominal
|
Target variable. Did it rain tomorrow? Yes or No
|
The Australian Weather dataset contains over 138,000 daily observations from January 2008 through to January 2017 from 49 Australian weather stations. Observations were drawn from numerous weather stations.
In completing Task 1 of Assignment 3 you will need to apply the business understanding, data understanding, data preparation, modelling and evaluation phases of the CRISP DM data mining process.
Task 1.1 Conduct an exploratory data analysis of the weatherAUS.csv data set using RapidMiner to understand the characteristics of each variable and the relationship of each variable to the other variables in the data set. Summarise the findings of your exploratory data analysis in terms of describing key characteristics of each of the variables in the weatherAUS.csv data set such as maximum, minimum values, average, standard deviation, most frequent values (mode), missing values and invalid values etc and relationships with other variables if relevant in a table named Task 1.1 Results of Exploratory Data Analysis for weatherAUS Data Set.
Hint: Statistics Tab and Chart Tab in RapidMiner provide a lot of descriptive statistical information and useful charts like Barcharts, Scatterplots etc. You might also like to look at running some correlations and chi square tests. Indicate in Task 1.1 Table which variables you consider to be the key variables which contribute most to determining whether it is likely to rain tomorrow.
Briefly discuss the key results of your exploratory data analysis and the justification for selecting your five top variables for predicting whether it is likely to rain tomorrow based on today's weather conditions. (About 250 words)
Task 1.2 Build a Decision Tree model for predicting whether it is likely to rain tomorrow based on today's weather conditions using RapidMiner and an appropriate set of data mining operators and a reduced weatherAUS.csv data set determined by your exploratory data analysis in Task 1.1. Provide these outputs from RapidMiner (1) Final Decision Tree Model process, (2) Final Decision Tree diagram, and (3) associated decision tree rules.
Briefly explain your final Decision Tree Model Process, and discuss the results of the Final Decision Tree Model drawing on the key outputs (Decision Tree Diagram, Decision Tree Rules) for predicting whether it is likely to rain tomorrow based on today's weather conditions and relevant supporting literature on the interpretation of decision trees (About 250 words).
Task 1.3 Build a Logistic Regression model for predicting whether it is likely to rain tomorrow based on today's weather conditions using RapidMiner and an appropriate set of data mining operators and a reduced weatherAUS.csv data set determined by your exploratory data analysis in Task 1.1. Provide these outputs from RapidMiner (1) Final Logistic Regression Model process and (2) Coefficients, and (3) Odds Ratios. Hint you will need to install the Weka Extension in RapidMiner, use W-Logistic Regression Operator for this Task 1.3 and you may need to change data types of some variables.
Briefly explain your final Logistic Regression Model Process, and discuss the results of the Final Logistic Regression Model drawing on the key outputs (Coefficients, Odds Ratios) for predicting whether it is likely to rain tomorrow based on today's weather conditions and relevant supporting literature on the interpretation of logistic regression models (About 250 words).
Task 1.4 You will need to validate your Final Decision Tree Model and Final Logistic Regression Model. Note you will need to use the X-Validation Operator; Apply Model Operator and Performance Operator in your data mining process models here.
Discuss and compare the accuracy of your Final Decision Tree Model with the Final Logistic Regression Model for whether it is likely to rain tomorrow based on today's weather conditions based the results of the confusion matrix, and ROC chart for each final model. You should use a table here to compare the key results of the confusion matrix for the Final Decision Tree Model and Final Logistic Regression Model (About 250 words).
Note the important outputs from your data mining analyses conducted in RapidMiner for Task 1 should be included in your Assignment 3 report to provide support for your conclusions reached regarding each analysis conducted for Task 1.1, Task 1.2, Task 1.3 and
Task 1.4. Note you can export the important outputs from RapidMiner as jpg image files and include these screenshots in the relevant Task 1 parts of your Assignment 3 Report.
Note you will find the North Text book a useful reference for the data mining process activities conducted in Task 1 in relation to the exploratory data analysis, decision tree analysis, logistic regression analysis and evaluation of the accuracy of the Final Decision Tree model and the Final Logistic Regression model.
Task 2
Research the relevant literature on how big data analytics capability can be incorporated into a data warehouse architecture. Note Chapter 2 Data Warehousing and Chapter 6 Big Data and Analytics of Sharda et al. 2014 Textbook will be particularly useful for answering some aspects of Task 2.
Task 2.1 Provide a high level data warehouse architecture design for a large stated owned water utility that incorporates big data capture, processing, storage and presentation in a diagram called Figure 1.1 Big Data Analytics and Data Warehouse Combined.
Task 2.2 Describe and justify the main components of your proposed high level data warehouse architecture design with big data capability incorporated presented in Figure 1.1 with appropriate in-text referencing support (about 750 words).
Task 2.3 Identify and discuss the key security privacy and ethical concerns for organisations within a specific industry that are already using a big data analytics and algorithmic approach to decision making with appropriate in-text referencing support (about 750 words).
Task 3
Scenario Dashboard
Los Angeles Police Department (LAPD) are responsible for enforcing law and order in the City of Los Angeles which is the cultural, financial, and commercial centre of Southern California. With a census-estimated 2015 population of 3,971,883, it is the second-most populous city in the United States (after New York City) and the most populous city in California. Located in a large coastal basin surrounded on three sides by mountains reaching up to and over 10,000 feet (3,000 m), Los Angeles covers an area of about 469 square miles (1,210 km2).
LAPD Crime Analytics Unit would like to have a Crime Events dashboard built with the aim of providing a better understanding of the patterns that are occurring in relation to different crimes across the 21 Police Department areas over time in the City of Los Angeles. In particular, they would like to see if there are any distinct patterns in relation to (1) types of crimes, (2) frequency of each type of crime across each of the 21 Police Department areas for years 2012 through to first quarter of 2016 based on the LACrimes2012-2016.csv data set. Note this is a large data set containing over 1 Million records. This Crime Events dashboard will assist LAPD to better manage and coordinate their efforts in catching the perpetrators of these crimes and be more proactive in preventing these crimes from occurring in the first place.
The LAPD Crime Analytics Unit wants the flexibility to visualize the frequency that each type of crime is occurring over time across each of the 21 Police Department areas/districts in the City of Los Angeles. They want to be able to get a quick overview of the crime data in relation to category of crimes, location, date of occurrence and frequency that each crime is occurring over time and then be able to zoom in and filter on particular aspects and then get further details as required.
LA Crimes Data Set Data Dictionary
|
variable name
|
type
|
Description
|
|
year_id
|
1. character
|
Original dataset id
|
|
date_rptd
|
2. date
|
Date crime was reported
|
|
dr_no
|
3. character
|
Count of Date Reported
|
|
date_occ
|
4. date
|
Date crime occurred
|
|
time_occ
|
5. date
|
Time crime occurred on a day
|
|
area
|
6. character
|
Area Code
|
|
area_name
|
7. character
|
Area geographical location
|
|
rd
|
8. character
|
Nearby road identifier
|
|
crm_cd
|
9. character
|
Crime type code
|
|
crm_cd_desc
|
10. character
|
Crime type description
|
|
Status
|
11. character
|
Status code
|
|
status_desc
|
12. character
|
Status outcome of crime
|
|
location
|
13. character
|
Nearby address location
|
|
cross_st
|
14. character
|
Nearby cross street
|
|
lat
|
15. numeric
|
Latitude of crime event
|
|
long
|
16. numeric
|
Longitude of crime event
|
|
year
|
17. numeric
|
Year of crime occurred
|
|
month
|
18. numeric
|
Month of crime occurred
|
|
day_of_month
|
19. numeric
|
Day of month crime occurred
|
|
hour_of_day
|
20. numeric
|
Hour of day crime occurred
|
|
month_year
|
21.
|
Month and year when crime occurred
|
|
day_of_week
|
22. character
|
Day of week crime occurred
|
|
weekday
|
23. character
|
Weekday/weekend classification for crime event
|
|
intersection
|
24. character
|
Occurred at an intersection
|
|
crime_classification
|
25. character
|
subjective binning of crimes
|
Task 3 requires a Tableau dashboard consisting of four crime event views of the LA Crimes 2012-2016 data set.
Task 3.1 Specific Crimes within each Crime Category for a specific Police Department Area and specific year
Task 3.2 Frequency of Occurrence for a selected crime over 24 hours for a specific Police Department Area
Task 3.3 Frequency of Crimes within each Crime Classification by Police Department Area and by Time
Task 3.4 Geographical (location) presentation of each Police Department Area for given crime(s) and year. Note for this task you will need to make use of the geo-mapping capability of Tableau Desktop.
You should briefly discuss the key findings for each of these four views in your Crimes Event Dashboard (about 60 words each and 250 words in total)
Task 3.5 Provide a rationale (drawing on relevant literature for good dashboard design) for the graphic design and functionality that is provided in your LAPD Crimes Event dashboard for the required four specified crime events views for Tasks 3.1, 3.2, 3.3 and 3.4 (About 750 words). Note Stephen Few is considered to be the Guru for good Dashboard Design and has wrote a number of books on this topic. Worth having a look at his website https://www.perceptualedge.com/about.php and in particular his examples of poorly designed dashboard views and his suggestions for better dashboard views.
For your Assignment 3 submission, you will need to submit your Task 3 Tableau workbook in .twbx format which will contain your dashboard, four views and the associated data set as a separate document together with your Assignment 3 Main Report in word docx format.
Report presentation writing style and referencing (worth 10 marks)
Presentation: use of formatting, spacing, paragraphs, tables and diagrams, introduction, conclusion, table of contents
Writing style: Use of English (Correct use of language and grammar. Also, is there evidence of spelling-checking and proofreading?)
Referencing: Appropriate level of referencing in text where required, reference list provided, used Harvard Referencing Style correctly