Many of the goals you accomplished in your individual course portfolio projects are incorporated into the Advanced Data Analytics capstone project including:
- Create a project proposal
- Demonstrate understanding of the form and function of Python
- Show how data professionals leverage Python to load, explore, extract, and organize information through custom functions
- Demonstrate understanding of how to organize and analyze a dataset to find the “story”
- Create a Jupyter notebook for exploratory data analysis (EDA)
- Create visualization(s) using Tableau
- Use Python to compute descriptive statistics and conduct a hypothesis test
- Build a multiple linear regression model with ANOVA testing
- Evaluate the model
- Demonstrate the ability to use a notebook environment to create a series of machine learning models on a dataset to solve a problem
- Articulate findings in an executive summary for external stakeholders
Project proposal
Salifort Motors is seeking a method to use employee data to gauge what makes them leave the company.
| Milestones | Tasks | PACE stages |
|---|---|---|
| 1 | Understand the business scenario and define the problem | Plan |
| 2 | Data exploration and data cleaning | Plan, Analyze |
| 3 | Determine which models are most appropriate | Analyze,Construct |
| 4 | Construct the model | Construct |
| 5 | Confirm model assumptions | Analyze, Construct |
| 6 | Evaluate model results | Analyze |
| 7 | Interpret results and share actionable steps with stakeholders | Execute |
Data Project Questions & Considerations
PACE: Plan Stage
Foundations of data science
- Who is your audience for this project?
- What are you trying to solve or accomplish? And, what do you anticipate the impact of this work will be on the larger business need?
- What questions need to be asked or answered?
- What resources are required to complete this project?
- What are the deliverables that will need to be created over the course of this project?
Get Started with Python
- How can you best prepare to understand and organize the provided information?
- What follow-along and self-review codebooks will help you perform this work?
- What are a couple additional activities a resourceful learner would perform before starting to code?
Go Beyond the Numbers: Translate Data into Insights
- What are the data columns and variables and which ones are most relevant to your deliverable?
- What units are your variables in?
- What are your initial presumptions about the data that can inform your EDA, knowing you will need to confirm or deny with your future findings?
- Is there any missing or incomplete data?
- Are all pieces of this dataset in the same format?
- Which EDA practices will be required to begin this project?
The Power of Statistics
- What is the main purpose of this project?
- What is your research question for this project?
- What is the importance of random sampling? In this case, what is an example of sampling bias that might occur if you didn’t use random sampling?
Regression Analysis: Simplify Complex Data Relationships
- Who are your stakeholders for this project?
- What are you trying to solve or accomplish?
- What are your initial observations when you explore the data?
- What resources do you find yourself using as you complete this stage? (Make sure to include the links.)
- Do you have any ethical considerations in this stage?
The Nuts and Bolts of Machine Learning
- What am I trying to solve?
- What resources do you find yourself using as you complete this stage?
- Is my data reliable?
- Do you have any additional ethical considerations in this stage?
- What data do I need/would I like to see in a perfect world to answer this question?
- What data do I have/can I get?
- What metric should I use to evaluate success of my business objective? Why?
Data Project Questions & Considerations
PACE: Analyze Stage
Get Started with Python
- Will the available information be sufficient to achieve the goal based on your intuition and the analysis of the variables?
Go Beyond the Numbers: Translate Data into Insights
-
What steps need to be taken to perform EDA in the most effective way to achieve the project goal?
-
Do you need to add more data using the EDA practice of joining? What type of structuring needs to be done to this dataset, such as filtering, sorting, etc.?
-
What initial assumptions do you have about the types of visualizations that might best be suited for the intended audience?
The Power of Statistics
- Why are descriptive statistics useful?
- What is the difference between the null hypothesis and the alternative hypothesis?
Regression Analysis: Simplify Complex Data Relationships
- What are some purposes of EDA before constructing a multiple linear regression model?
- Do you have any ethical considerations in this stage?
The Nuts and Bolts of Machine Learning
- What am I trying to solve? Does it still work? Does the plan need revising?
- Does the data break the assumptions of the model? Is that ok, or unacceptable?
- Why did you select the X variables you did?
- What are some purposes of EDA before constructing a model?
- What has the EDA told you?
- What resources do you find yourself using as you complete this stage?
- Do you have any ethical considerations in this stage?
Data Project Questions & Considerations
PACE: Construct Stage
Get Started with Python
- Do any data variables averages look unusual?
- How many vendors, organizations or groupings are included in this total data?
Go Beyond the Numbers: Translate Data into Insights
- What data visualizations, machine learning algorithms, or other data outputs will need to be built in order to complete the project goals?
- What processes need to be performed in order to build the necessary data visualizations?
- Which variables are most applicable for the visualizations in this data project?
- Going back to the Plan stage, how do you plan to deal with the missing data (if any)?
The Power of Statistics
- How did you formulate your null hypothesis and alternative hypothesis?
- What conclusion can be drawn from the hypothesis test?
Regression Analysis: Simplify Complex Data Relationships
- Do you notice anything odd?
- Can you improve it? Is there anything you would change about the model?
The Nuts and Bolts of Machine Learning
- Is there a problem? Can it be fixed? If so, how?
- Which independent variables did you choose for the model, and why?
- How well does your model fit the data? (What is my model’s validation score?)
- Can you improve it? Is there anything you would change about the model?
- Do you have any ethical considerations in this stage?
Data Project Questions & Considerations
PACE: Execute Stage
Get Started with Python
-
Given your current knowledge of the data, what would you initially recommend to your manager to investigate further prior to performing an exploratory data analysis?
-
What data initially presents as containing anomalies?
-
What additional types of data could strengthen this dataset?
Go Beyond the Numbers: Translate Data into Insights
-
What key insights emerged from your EDA and visualizations(s)?
-
What business recommendations do you propose based on the visualization(s) built?
-
Given what you know about the data and the visualizations you were using, what other questions could you research for the team?
-
How might you share these visualizations with different audiences?
The Power of Statistics
-
What key business insight(s) emerged from your A/B test?
-
What business recommendations do you propose based on your results?
Regression Analysis: Simplify Complex Data Relationships
-
To interpret model results, why is it important to interpret the beta coefficients?
-
What potential recommendations would you make to your manager/company?
-
Do you think your model could be improved? Why or why not? How?
-
What business recommendations do you propose based on the models built?
-
What key insights emerged from your model(s)?
-
Do you have any ethical considerations at this stage?
The Nuts and Bolts of Machine Learning
-
What key insights emerged from your model(s)?
-
What are the criteria for model selection?
-
Does my model make sense? Are my final results acceptable?
-
Were there any features that were not important at all? What if you take them out?
-
Given what you know about the data and the models you were using, what other questions could you address for the team?
-
What resources do you find yourself using as you complete this stage?
-
Is my model ethical?
-
When my model makes a mistake, what is happening? How does that translate to my use case?