From Industry Problem to Analytics Solution
Today's Goal: Leave this workshop with significant progress on your Assessment 1 Literature Review
Assessment 1 Due: Week 5 (2 weeks from today)
Assess where you are in your Assessment 1 journey and identify gaps
Understand real-world data analytics challenges and how they affect YOUR project
Match your business problem to appropriate analytics methods and tools
Make tangible progress on your literature review with peer feedback
By the end of today: You should have your industry, 3 business problems, and at least 5 relevant sources identified.
Week 1-2
Industry selection
Week 3 (NOW)
3 problems + Sources
Week 4
Draft + Visualisations
Week 5
Submit
Answer honestly - this helps identify what you need to focus on today.
If you selected A or B: Today's workshop is critical for you. Use the activities to make rapid progress.
Find data → Clean it → Analyse → Done!
📊 → 🧹 → 📈 → ✅
Find data → Wrong format → Clean → Missing values → Re-clean → Tool issues → Finally analyse...
📊 → ❌ → 🧹 → ❓ → 🔄 → 🛠️ → 📈
Why this matters for Assessment 1: Understanding these challenges helps you write a realistic methodology section and choose appropriate data sources.
| Activity | Time |
|---|---|
| Data cleaning & organizing | 60% |
| Collecting data sets | 19% |
| Mining data for patterns | 9% |
| Refining algorithms | 4% |
| Building training sets | 3% |
| Other | 5% |
Source: CrowdFlower Data Science Survey
When evaluating data sources, consider:
Assessment Tip: Address data quality in your "Data Sources" section!
Remember: When discussing data sources in your assessment, mention potential data quality issues and how they might be addressed.
Your methodology must clearly state which type(s) of analytics you will use.
Question: "What happened?"
Methods: Summary statistics, data aggregation, visualisation, dashboards
Example: Analysing last year's sales by region
Question: "What could happen?"
Methods: Regression, classification, forecasting, machine learning
Example: Predicting customer churn probability
Question: "What should we do?"
Methods: Optimisation, simulation, decision analysis
Example: Recommending optimal pricing strategy
Increasing complexity and business value
Your Assessment 1 must briefly outline the methodologies you will explore.
| Tool | Best For | Skill Level | Your Project? |
|---|---|---|---|
| Tableau / Power BI | Visualisation, dashboards, descriptive analytics | Beginner-Intermediate | Visualisations required |
| Python (Pandas, Scikit-learn) | Data manipulation, ML, predictive analytics | Intermediate-Advanced | If doing predictive |
| R | Statistical analysis, visualisation | Intermediate-Advanced | Statistical focus |
| Excel | Basic analysis, pivot tables | Beginner | Simple analyses |
| SQL | Data extraction, database queries | Beginner-Intermediate | If data is in databases |
Assessment 1 Requirement: You must upload your visualisation file (Tableau, Power BI) to the file Dropbox. Plan accordingly!
Creating complex models when simple analysis would suffice.
Symptom: Spending weeks on deep learning when descriptive statistics answer the question.
Solution: Start simple. Add complexity only if needed.
Choosing tools because they're "cool" rather than appropriate.
Symptom: Using neural networks for a dataset of 100 records.
Solution: Match tools to problem complexity and data size.
Creating brilliant analysis that no one understands.
Symptom: Technical jargon overwhelming stakeholders.
Solution: Focus on key findings and actionable insights.
Trying to solve every problem at once.
Symptom: Assessment keeps growing beyond 1000 words.
Solution: Focus on ONE clear business question.
Complete this mapping for YOUR Assessment 1 project:
| Component | Your Response |
|---|---|
| Industry: | |
| Business Problem 1: | |
| Analytics Type: | ☐ Descriptive ☐ Predictive ☐ Prescriptive |
| Business Problem 2: | |
| Analytics Type: | ☐ Descriptive ☐ Predictive ☐ Prescriptive |
| Business Problem 3: | |
| Analytics Type: | ☐ Descriptive ☐ Predictive ☐ Prescriptive |
| Primary Tool(s): |
Assessment 1 requires you to "evaluate the types of data sources available."
| Industry | Example Data Sources |
|---|---|
| Healthcare | AIHW, Medicare statistics, hospital discharge data, clinical trials |
| Retail | Kaggle datasets, company reports, ABS retail trade data |
| Finance | ASX data, Yahoo Finance, RBA statistics, company filings |
| Government | data.gov.au, ABS, state government open data portals |
For each potential data source, complete this evaluation:
| Criteria | Data Source 1 | Data Source 2 |
|---|---|---|
| Name/URL: | ||
| Accessibility (1-5): | ||
| Data Quality (1-5): | ||
| Relevance (1-5): | ||
| Key Variables: | ||
| Limitations: | ||
| Will use for Assessment? | ☐ Yes ☐ No ☐ Maybe | ☐ Yes ☐ No ☐ Maybe |
Assessment 1 requires at least 10 relevant, credible references.
Use the printed handout to complete this activity.
This structured exercise will help you:
Important: Complete the handout thoroughly. By the end, you should have identified at least 2 potential industry-dataset combinations to explore further.
Assessment 1 requires you to generate a unique business question for ONE of your problems.
A good business question is:
| Business Problem | Weak Question | Strong Question |
|---|---|---|
| Customer churn | "Why do customers leave?" | "Which customer behaviors in the first 30 days predict churn within 6 months?" |
| Hospital readmissions | "How can we reduce readmissions?" | "What patient characteristics and discharge factors predict 30-day readmission for cardiac patients?" |
| Sales forecasting | "What will sales be?" | "How do seasonal patterns and promotional activities influence weekly sales volume by product category?" |
You must explain why your analysis is original given existing research.
"While Smith (2022) examined customer churn in US telecommunications, no study has applied these methods to the Australian market with its unique regulatory environment."
"Previous research focused on demographic factors; this analysis incorporates social media sentiment data not available in earlier studies."
Tip: Your literature review should naturally lead to identifying this gap. If you can't find a gap, your literature review may not be comprehensive enough.
Critical Requirement: You need at least THREE relevant visualisations from YOUR dataset. The Tableau/Power BI file must be uploaded to the file Dropbox.
Purpose: Show data overview/distribution
Examples: Histogram, bar chart, pie chart
Demonstrates: You understand your data
Purpose: Show relationships/trends
Examples: Scatter plot, line chart, heatmap
Demonstrates: Insight into patterns
Purpose: Support your business question
Examples: Depends on your question
Demonstrates: Data can answer your question
By next week: Have your dataset downloaded and create at least one draft visualisation to bring to class.
Using the framework from Slide 20, draft your unique business question:
| Selected Business Problem: | |
| Draft Question (v1): |
| Refined Question (v2): |
Your literature review must communicate complex ideas clearly.
"Stakeholders need the key findings and action items. Save the technical details for the appendix."
— Andrew Seitz, Senior Data Analyst, Snowflake
By next week, you should have:
Week 4 Workshop: Bring your draft literature review for peer feedback. We will workshop your visualisations and methodology sections.
Descriptive, Predictive, or Prescriptive - match to your business problem
Match tools to your skills, data, and timeline. Tableau/Power BI required for visualisations.
Consider accessibility, quality, and relevance before committing.
2 weeks to submission. Draft early, revise often.
Questions? Use the remaining workshop time for facilitator consultation.