Main QuestionI will attach an example of how the power point should look like!!! (MIS TOPIC 8, MIS Topic 7, MIS Topic 6, MIS topic 5 are all examples… I do think one of the examples do not include speaker notes… BUT I NEED SPEAKER NOTES ON MY PRESENTATION!!!Create a PowerPoint presentation of 25 slides, not including
the title and references slides, that summarizes your project and
integrates the work you have done on the project in Topics 1-7.(I am going to attach our most resent paper that shoud answer topics 1-6. You just completed Topic 7. The Power Point needs to include the following..) 1. Speaker Notes… ( The speaker notes can come directly out of the paper, they just have to fit with the slide.2. Color- The presentation needs to be visually inspiring. You have to use lots of color3. Bullet points- NO COMPLETE SENTENCESAssume that you are delivering this presentation to the senior
leadership in an organization. Therefore, please be sure to create a
professional presentation.Note that you are summarizing the information that is specifically
tailored for senior management. Be thorough, but also remember to be
concise. Organize the PowerPoint presentation according to the following:Business Problem and supporting information (based on your
work in Topic 1).Analytics Problem and supporting
information (based on your work in Topic 2).Results of the
Data Needs Identification and Data Acquisition and associated
supporting information (based on your work in Topic 3).Data cleansing and Data Summary and associated supporting
information (based on your work in Topic 4).Model Building
and Model Validation (based on your work in Topics 5 and 6).Model Deployment and Model Life Cycle (based on your work in
Topic 7).Specific future challenges and concluding
recommendations to senior management regarding your model.In the “Notes” section of each slide, include your talking
points. These notes should provide information that would be verbally
conveyed when delivering the presentation in-person.Refer to the resources, “Creating Effective PowerPoint
Presentations,” located in the Student Success Center, for
additional guidance on completing this assignment in the appropriate style.While APA style is not required for the body of this assignment,
solid academic writing is expected, and documentation of sources
should be presented using APA formatting guidelines, which can be
found in the APA Style Guide, located in the Student Success Center.This assignment uses a rubric. Please review the rubric prior to
beginning the assignment to become familiar with the expectations for
successful completion.You are required to submit this assignment to LopesWrite. Please refer
to the directions in the Student Success Center.QUESTION 1- The 2016-17 Miami Heat began the season with an 11-30 record before a magical turnaround that saw the team’s final record end up at 41-41. At the midway point of the season, they had played 18 home games and 23 away games. Assume that, in that season, the Heat had a 45% chance to win a home game (regardless of previous results, injury, fatigue, or other factors) and a 30% chance to win an away game (again, regardless of all factors). Calculate the following probabilities:1) That the Heat would win at most 11 of their first 41 games; and2) That the Heat would win at least 30 of their final 41 gamesBe thorough and explicit with your explanation and presentation of your approach to this problem.Question 2- (my computer does not run SQL very well because its a mac… I am going to try to get you this file… If not just describe to the best of your knowledge how to set this problem up to be answered… For example, what commands to use in SQL to get this type of information up…)Assume Real Madrid defines “Game Goals” for an upcoming game — separated into offensive and defensive goals — as the average of the metric in their own wins and their opponent’s losses for each of basketball’s Four Factors (effective field goal percentage, turnover percentage, rebounding percentage, and free throw attempt rate).Assume they define their post-game Goal Achievement Average, or “GAA”, as the average of the differences between binary values indicating each goal’s completion and the pre-game probability of the corresponding goal’s completion.Download and import the SQL database found at tinyurl.com/PHX2019DataChallenge into your preferred SQL Database Client, then answer the following questions.1) Calculate Real Madrid’s Game Goals for their 25th game of the Euro League season against Fenerbahce Istanbul.2) For each Game Goal, estimate the pre-game probability that Real Madrid would accomplish the goal in the game against Fenerbahce.3) Provide Real Madrid’s post-game GAA for the game.4) Provide a SQL query that calculates the Game Goals for every team in every game so far in the Euro League season. Your query should return 400 records.State your assumptions, observations, and modeling decisions concisely in your writeup along with your answers to questions 1-3.Question 3-Statistical and predictive draft models are a crucial component to any analytics department’s value added. With any draft model being deeply tied to past results while also aiming to predict future ones, the recent evolution in style of play in the NBA presents an interesting problem for draft modeling.Comment on how you would approach the need to account for the aforementioned recent evolution in a draft model of your own. Assume you are limited to all publicly available data and video (amateur and NBA) of past and current players. Discuss any additional data points you would seek to gather or calculate, advanced techniques you would employ, and any other modeling considerations.Be concise and clear in your answers, combining an understanding of the impact that recent changes have had and will have on the NBA along with detailed and insightful explanation of technical modeling nuances, challenges, and solutions.
mis_690_topic_8_1_.pptx
mis_690_topic_7_1_.pptx
mis_690_topic_5.pptx
mis_690_topic_6.pptx
rubric_print_format_9_.xlsx
Unformatted Attachment Preview
MIS 690 Capstone
TOPIC 8: Capstone Recommendations and Reporting
Grand Canyon University
Dr. Randy Bulriss
Agenda
1) Group Project Show Case – 120 mins
2) Dinner (Potluck in Room) 30 mins
3) Class Retrospective and Close – 30 mins
2
R. Bulriss Grand Canyon University
1/11/18
Lecture Objectives
1) Group Presentations
Key Result: Process retrospective and process improvement.
2) Topic 8
expectations and Role Model Example.
Key Result: Enable Individual Presentation Success
3
R. Bulriss Grand Canyon University
1/11/18
MIS 690 Capstone Objectives
Week 1 Objectives: Business Problem Statement
Week 2 Objectives: Analytics Problem Statement
Validate an analytics model for a given project.
Week 7 Objectives: Model Deployment and Model Life Cycle
Select an appropriate analytics model for a given project.
Build an analytics model for a given project.
Week 6 Objectives: Model Validation
Cleanse data for analysis.
Summarize data behavior to inform subsequent decision making.
Week 5 Objectives: Methodology Approach and Model Building
Determine data needs to solve an analytics problem.
Week 4 Objectives: Benchmark – Data Cleansing and Data Summary
Analyze a business problem using specific analytics information.
Compose an analytics business problem statement.
Week 3 Objective: Data Needs Identification and Data Acquisition
Analyze a business problem.
Compose a business problem statement.
Describe model deployment and aspects of the model life cycle.
Articulate specific issues related to deploying an analytics model and its lifecycle.
Week 8 Objectives: Benchmark – Capstone Recommendations and Reporting
Recommend analytics-based solutions for business problems.
Use results of data analysis to identify and make recommendations for specific analytics solutions that could be
deployed to solve the identified business problem.
Communicate data analysis results to organizational stakeholders using visualization techniques.
4
R. Bulriss Grand Canyon University
1/11/18
MIS 690 Capstone Assignments
Week 8 Objectives: Benchmark – Capstone Recommendations
Recommend analytics-based solutions for business problems.
Use results of data analysis to identify and make recommendations for specific
analytics solutions that could be deployed to solve the identified business problem.
Communicate data analysis results to organizational stakeholders using visualization
techniques.
5
Details: INDIVIDUAL
Create a PowerPoint presentation of 15 to 25 slides, not including the title and references slides, that summarizes your project and
integrates the work you have done on the project in Topics 1-7.
Assume that you are delivering this presentation to the senior leadership in an organization. Therefore, please be sure to create a
professional presentation.
Note that you are summarizing the information that is specifically tailored for senior management. Be thorough, but also remember to be
concise. Organize the PowerPoint presentation according to the following:
Business Problem and supporting information (based on your work in Topic 1).
Analytics Problem and supporting information (based on your work in Topic 2).
Results of the Data Needs Identification and Data Acquisition and associated supporting information (based on your work in Topic 3).
Data cleansing and Data Summary and associated supporting information (based on your work in Topic 4).
Model Building and Model Validation (based on your work in Topics 5 and 6).
Model Deployment and Model Life Cycle (based on your work in Topic 7).
Specific future challenges and concluding recommendations to senior management regarding your model.
In the “Notes” section of each slide, include your talking points. These notes should provide information that would be verbally conveyed
when delivering the presentation in-person.
Refer to the resources, “Creating Effective PowerPoint Presentations,” located in the Student Success Center, for additional guidance on
completing this assignment in the appropriate style
GROUP Summary Retrospective Presentations in CLASS
R. Bulriss Grand Canyon University
1/11/18
Role Model 1
Determining Sales Prices Using Statistics
6
R. Bulriss Grand Canyon University
1/11/18
Determining Sales Prices Using Statistics
Capstone Project
MIS-690| Grand Canyon University| Jacobus DeBruyn
Project Objectives
▪
Provide a tool to determine the sales price for a new sales generated.
▪
Pricing decisions should be data driven.
▪
Flexibility to react to changes in market conditions.
▪
Use of descriptive statistics to identify outliers to the data set.
▪
Develop procedures to minimize future outliers.
▪
Create a repeatable process.
Table of Contents
▪
Business Problem
▪
Analytics Problem
▪
Data Needs Identification and Data Acquisition
▪
Data Cleansing and Data Summary
▪
Model Building and Model Validation
▪
Model Deployment and Model Life Cycle
▪
Future Challenges and Recommendations
Business Problem
Background
▪
The company operates in a seasonal market.
▪
During peak periods, capacity is limited.
▪
Off-peak seasons, capacity is available.
▪
Aggressive pricing strategy is needed to win business.
▪
Quick turnaround needed between pricing request and quote delivery.
▪
Pricing should give the company a competitive advantage.
Business Problem
▪
The company vision is to be a $2 billion-dollar company by 2020.
▪
Product profitability of 12% on all products.
▪
A 3% profitable growth year-over-year.
▪
The goal is to achieve this through strengthened pricing.
▪
A robust tool is needed to assist with pricing decisions.
Analytics Problem
Stakeholders
▪
Stakeholder
“A person or group of people who can affect or be affected by a given project.”
(Piscopo, n.d.).
▪
Project Stakeholders
– Billing and collections team
– Pricing team
– Sales team
– Factory cost teams
Analytics Problem
▪
Multiple regression analysis will be used for the analysis.
▪
The relationship between the Sales Unit Price and the following variables will be
evaluated:
– Quarters,
– Wire Type,
– Variable Unit Cost,
– Fixed Unit Cost,
– Sales Quantity.
▪
Data for the analysis sourced and extracted from the ERP system.
Multiple Regression Analysis
▪
Predict the value of a variable based on the value of two or more other variables.
▪
Assumptions:
– Dependent variable measured on a continuous scale,
– Two or more independent variables,
– Independence of observations,
– Linear relationship,
– Homoscedasticity,
– Do not show multicollinearity,
– No significant outliers (Laerd Statistics, n.d.).
Data Needs Identification and Data Acquisition
Data Acquisition Process
▪
Data source: transactional ERP system
▪
Functional Areas
– Sales and Distribution
– Cost Accounting
▪
Data consolidation and cleanup in Microsoft Excel
– Blank values
– Missing data
– Outliers (maximum unit sales price)
Data Description
Data Identification and Acquisition
Data Source Tables
Variables Identified
Data Cleansing and Data Summary
Data Quality & Missing Data
Data Cleansing – Round 1
Data Cleansing – Round 2
1.
Identify and remove missing data.
1.
Use round 1 data as a base.
2.
Add additional calculations to the data set.
2.
Correct Region mapping issue.
3.
Analyze Sales by Sales Region.
3.
Remove anomalies using the Package Guidance.
4.
Analyze Unit Sales by Package and Region.
4.
Remove Wire Application as a variable.
5.
Perform Trend Analysis.
6.
Calculate a Statistical Summary for each Package.
Data Relationship
Data Trends
Data Outliers
Data Table
Scatter Plot
Data Segmentation
Model Building and Model Validation
First Multiple Regression Analysis
▪
Assumed α = 0.05.
▪
Independent variables have a
statistically significant impact if pvalues < 0.05.
▪
Variables related to quarters and
regions > 0.05.
▪
Variables removed from the analysis.
Second Multiple Regression Analysis
▪
Multiple regression analysis rerun
with excluded variables.
▪
Remaining dependent variables have
a statistically significant impact.
▪
Multiple Regression Model can be
used for predictive analysis.
Model Conclusion
▪
Weibull distribution.
▪
Independent variables:
– Unit Variable Cost,
– Unit Fixed Cost.
▪
Variables statistically significantly
predicted Unit Sales Price, F(2, 5882)
= 2509.057, p < .0005, R2 = .460.
▪
Both independent variables added
statistically significantly to the
prediction, p < .05.
Validation
▪
Splitting the data set into training and validation data sets.
▪
All data extracted were transformed to build the multiple regression analysis
model.
▪
An additional data set needed to test the model.
▪
Data for Quarter 1, 2015 was used for validation.
▪
The same parameters were used to extracted and transformed the new data set.
▪
The new data set was used for comparison multiple regression analysis model.
Validation Steps
▪
Step 1: Dependent Variable
▪
Step 2: Excluded Variables
▪
Step 3: Coefficients and ANOVA table
▪
Step 4: Correlation Coefficient (R)
▪
Step 5: Coefficient of Determination (R2)
▪
Step 6: Independent Variable Significance
▪
Step 7: Normal Probability
Model Deployment and Model Life Cycle
Model Life Cycle
Source: Lorica (2013).
Model Deployment Cost
▪
All stakeholders were trained.
▪
License fee (yearly), fixed for next 3
years.
▪
Future models will only require
consulting fees.
▪
Internal resources to support
enhancements.
Deployment Timeline and Proposed Tasks
Future Challenges and Recommendations
Future Challenges and Recommendations
▪
Model recalibration and maintenance.
▪
Evaluate use of the multiple regression analysis.
▪
Validation of model quality.
▪
Training plan for new users.
▪
Visualization of data.
▪
Scalable system.
▪
Continuous management support.
Questions?
Review
▪
On a daily basis new pricing requests are received from customers.
▪
It is important to respond as fast as possible to the request.
▪
Delayed negotiations can result in a loss of sales.
▪
Pricing is a complex discipline.
▪
Data drive decisions allows for fast response to pricing requests.
▪
Decisions are based on historical data.
▪
Data is working for the user to assist in the decision making process.
References
▪
Kostoulas, A. (2014, December 8). Dependent and Independent Variables, Using
SPSS, and Minding One’s Manners [Web log]. Retrieved from
https://achilleaskostoulas.com/2014/12/08/dependent-and-independent-variablesusing-spss-and-minding-ones-manners/
▪
Laerd Statistics. (n.d.). Multiple Regression Analysis using SPSS Statistics. Retrieved
from https://statistics.laerd.com/spss-tutorials/multiple-regression-using-spssstatistics.php
▪
Lorica, B. (2013, July 14). Data scientists tackle the analytic lifecycle. Retrieved from
https://www.oreilly.com/ideas/data-scientists-and-the-analytic-lifecycle
References
▪
Piscopo, M. (n.d.). What is a Stakeholder? How to Identify, Analyze and Manage
Project Stakeholders. Retrieved from
http://www.projectmanagementdocs.com/blog/what-is-astakeholder.html#axzz51qd8OtlX
Role Model 2
Improving Customer Satisfaction
43
R. Bulriss Grand Canyon University
1/11/18
IMPROVING CUSTOMER SATISFACTION
CAPSTONE PROJECT
SEPTEMBER 20, 2017
MIS 690 STUDENT
WHAT IS THE PROBLEM?
• What can we do about it?
• Improve the outage experience
• How can the outage experience be improved?
• Collect feedback from customers who recently experienced an outage
• Analyze the data to determine what impacts customer satisfaction
•
•
Use modeling techniques to uncover the biggest influencers of satisfaction
Results will serve as key performance metrics to proactively manage and improve over time
WHY SHOULD WE FIX THE PROBLEM?
• Why should we improve the outage experience and ultimately customer
satisfaction?
• Ease the strain on the call center during outages
• Reduce the number of customer complaints filed with the Better Business Bureau and/or the SRP
Ombudsman
• Decrease the chances of costly and time consuming litigation
• Social responsibility to provide reliable power and good customer service
• Build trust and good faith with the community
• Help customers better prepare and cope with outages
• Meet corporate goal and trigger financial reward for employees
HOW CAN WE DEFINE CUSTOMER SATISFACTION?
• Main objective is to improve customer satisfaction
• Four areas of satisfaction
•
•
•
•
SRP’s power quality and reliability (PQR)
Customer service representative (CSR)
Outage experience
Overall satisfaction with SRP
• Understanding these four areas is key in providing actionable, implementable results
• Provides more granular information for improved decision making
• Four analytical models will be created to uncover the drivers of each area of satisfaction
HOW DO WE OBTAIN THE DATA?
• Use data obtained through historical outage management tracking study
• Data are collected on a weekly basis throughout the year through an online
survey hosted by a third-party vendor
• Data used for this project spans from June 2016 to April 2017
• Survey respondents must have experienced an outage in the two weeks prior
to receiving an email survey invite
• Asked to evaluate their outage experience and the outage reporting channel used:
• Electronic notifications (text and email)
• Conversational Interactive Voice Response (CIVR)
• Customer Service Representative (CSR)
WHAT ANALYTICAL TECHNIQUE SHOULD BE USED?
• Analytical goals:
• Determine relationships between attributes and satisfaction
• Determine importance or level of impact of attributes on satisfaction
Stepwise
Multiple
Regression
WHAT DATA WERE USED FOR EACH MODEL?
PQR Satisfaction
• Providing power during extremely hot
temperatures
• Ensuring an adequate supply of power
• Providing timely restoration of power
• Having a minimal number of power
outages throughout the year
• Frequency of outages
• Duration of outages
CSR Satisfaction
• Ease of getting through to a CSR
• CSR was helpful
• CSR was courteous
• CSR showed concern for question
• CSR had sufficient knowledge
• CSR listened
• CSR was able to answer the question on
first call
• CSR provided information that was clear
and easy to understand
• Frequency of outages
• Duration of outages
WHAT DATA WERE USED FOR EACH MODEL?
Outage Experience
Satisfaction
•Easy to contact during outage
•Providing information on outages
•Satisfaction with CIVR
•Satisfaction with text and/or email electronic notifications
•Satisfaction with CSR
•Satisfaction with amount of time it took to restore power
•Time of power restoration
•Received a call back when power was restored
•Satisfaction with overall communications about outage
•Satisfaction with receiving all outage information needed
•Frequency of outages
•Duration of outages
•Type of information received
Overall Satisfaction
with SRP:
All data used for
previous 3 models
MORE INFORMATION ON DATA INPUTS
• Two types of data used in this project
• Discrete
• Majority of data are rating questions whose answer choices are
on a scale of 1 to 10 or 1 to 5 (1=lowest rating; 5/10=highest
rating)
• Categorical questions (i.e., Yes/No/Don’t know)
• Occurrence counts (i.e., number out outages in the past 6 months)
• Continuous
• Only one – duration of outage
ARE THE DATA AND METHODOLOGY SUFFICIENT TO
ANSWER THE PROBLEM?
• Yes!
• Historical project where the data are sourced from designed around the goal of
assessing and improving outage experience and satisfaction
• Survey questionnaire updated regularly to reflect the current outage experience
• Stepwise multiple regression analysis will pinpoint those attributes that have the biggest
impact on satisfaction
DATA CLEANING
• Removed invalid responses from the variables
• Removed outliers from continuous/numeric variables
• Visual assessment using box and whisker plots
• Comparison of mean, median and standard deviation
• Assessment of missing values
• New questions added mid-year to the survey
• Removal of invalid responses
• Survey design
WHAT PATTERNS EMERGED?
• Most scores for the rating
questions fall on the upper, or
more satisfied, end of the scale
• Respondents are generally satisfied
with outage experience/interactions
• In general, the longer the outage,
the lower the satisfaction
ARE THERE SEASONALITY
EFFECTS?
• Longer and more outages during the warmer
months of the year
• More outages experienced, the lower the
satisfaction rating
• Lowest satisfaction during summer months
• Keep mind: satisfaction at no point is low – just
lower during certain months
• CSR satisfaction doesn’t show a clear pattern
• Other factors likely exerting more influence
WHAT IS STEPWISE MULTIPLE REGRESSION?
• Four models built using stepwise multiple regression
•
•
•
•
Determines the most important predictors of a dependent variable – satisfaction
Improves model performance by adding only the best predictors into the model
Reduces manual input by automatically adding the variables
SPSS produces metrics to evaluate model performance
• Adjusted R2
•
•
Determines how well the model fits the data
•
•
Generally, the higher the better (ranges from 0% to 100%)
Calculates how much variation in the dependent variable (satisfaction) can be explained by the input
variables
Adjusted for the number of input variables entered into the model
WHAT INFLUENCES PQR SATISFACTION?
• Most important predictors:
• Having a minimal number of power outages
• Ensuring an adequate supply of power
• Providing timely restoration
• 78% of the variation in PQR satisfaction
can be explained by these three predictor
variables
• Indicates good model performance
WHAT INFLUENCES CSR SATISFACTION?
• Most important predictors:
• Being helpful
• Answering the question on the first call
• 67% of the variation in CSR satisfaction can be
explained by these two predictor variables
• Indicates good model performance
WHAT INFLUENCES OUTAGE EXPERIENCE
SATISFACTION?
• Most important predictors:
•
•
•
•
•
•
Satisfaction with the amount of time it took to restore power
Providing information on outages
Satisfaction with the overall outage communications
Being easy to contact when outages occur
Satisfaction with the CSR
Frequency of outages
• 72% of the variation in outage experience satisfaction can be explained by these
six predictor variables
• Indicates good model performance
WHAT INFLUENCES OVERALL SATISFACTION?
• Most important predictors:
•
•
•
•
•
•
•
Satisfaction with PQR
Providing timely restoration of power after an outage
Providing information on outages
Overall outage experience
Ensuring an adequate supply of power
Satisfaction with overall outage communications
CSR being courteous
• 69% of the variation in overall satisfaction can be explained by these seven
predictor variables
• Indicates good model performance
WHEN AND HOW WILL WE USE THE MODELS?
May
August
November
FY Start
May 1
February
May
FY End
April 30
• Manual deployment of models biannually
• Ensures a large enough sample of survey respondents
• Allows close monitoring of model performance and accuracy
• Not necessary to operationalize the models
•
•
•
•
Simple to run
Larger time span between runs
Location of data storage
Nature of the project
Quarterly Report
Model Deplo …
Purchase answer to see full
attachment
