Week 4: Question-Focused Analysis

Agenda

Identify a question

Develop a hypothesis

Tell the story with your data

Brainstorm project ideas

Identify a question

A question might be driven by a business decision or goal

• We need to cut costs. Can we reallocate labor more efficiently?

• We want to increase profits. Can we increase marketing efforts on profitable product lines?

Refine the question

Can we answer the question with the available data?

Is the question specific enough?

Will the answer explain a phenomenon?

1. Can we answer the question with the available data? If not, can we adjust the question or do we need to collect more data

2. specific enough? Is climate change impacting the agricultural sector? More specific: How are CO corn growers adapting to water shortages, which are expected to increase because of climate change?

3. explain phenomenon? Does the question lend itself to explanations that help form testable hypotheses? In the previous question, are growers adopting more efficient irrigation technologies? Are they selling farm land and exiting agriculture?

Exercise

Are these questions answerable with the supermarket_sales dataset? why?

1. We need to cut costs. Can we reallocate labor more efficiently?

2. We want to increase profits. Can we increase marketing efforts on profitable product lines?

Develop a hypothesis

What do you think the answer is? Why?

The Why part of your answer will help guide your analysis of the data

Let the question drive the analysis

Think through how you can analyze the data to answer your question

The evidence should support your conclusions

Example Analysis

We have a limited budget for a promotion campaign.

Question: Which product(s) have a high customer rating by members? What day of the week do members buy those products?

What kind of visual would help you answer part 1? and part 2?

Example Analysis cont.

Which product(s) have a high customer rating by members?

What day of the week do members buy those products?

Example Analysis cont.

What else might you investigate during your exploratory data analysis (EDA)?

Does this vary by city/branch?

Keep a record of your exploration

Keep a notebook as you explore your dataset

How did you find your insight? Especially important for Tableau

Your process may help you tell your story

Walking the audience through your thought process can help them understand your analysis. Here is my question… Then I thought, what does the data look like over time or across cities… Then, I noticed this pattern on Wednesdays…

Know when to quit

Data analysis can be fun

You may be tempted to “check one more thing”

Stay focused on your question and quit when you have answered it

Tell the story with data

Develop a narrative

Create effective visuals

Let’s continue with our example analysis

For the mini project, the final deliverable will be a report rather than a presentation. You can still use many of the story telling techniques like having a narrative and effective visuals

Story overview

Start with the question¹

Characters and setting

Preliminaries

Answer (or insight)

Next steps

1. Sometimes you won’t have a clear question. You are communicating an insight and you need to get the audience to care.

Question

If given a question, review it and transition into setting¹

In our example analysis, we were given the questions:

• Which product(s) have a high customer rating by members?

• What day of the week do members buy those products?

1. If not, you may need to develop the setting a bit

Characters and context

Who are the characters?

• Customers: the members and non members

What is the setting?

• Sales in Q1 2019 at three stores

• Transaction level data where customers rate the product and buying experience after a purchase

Preliminaries

You will often need to set up your main point with supporting evidence (effective visuals)

A common strategy is to start with the big picture and drill down to the insight or answer to the question

Where could you start with the example analysis?

• Sales by product category

How would you drill deeper?

Key insight or answer

Your key insights is the most important part of your story

Make sure the visual is effective (i.e., does it clearly illustrate the point?)

Members rate health and beauty products highest among categories

Figure 1. Average transaction rating by members across product category in 2019 Q1.

Members spend more on health and beauty products on the weekend

Figure 2. Average spending on health and beauty per transaction by members by day of week in 2019 Q1.

Next steps

Make recommendations (if appropriate)¹

Is any additional analysis needed? Briefly describe

Example analysis:

• Design promotion for health and beauty products on the weekend

• Conduct market survey to better understand why members purchase these items on the weekend

1. It may not be appropriate for you to make direct recommendations. Try to frame the discussion.

Brainstorm project ideas

Understanding customer behavior: analyze customer behavior and preferences, such as the most popular product lines and the average purchase frequency.

Sales and revenue analysis: analyze sales and revenue trends, such as the average invoice value and the contribution of different branches to the overall revenue.

Customer segmentation: segment customers based on demographics, such as gender and customer type, to understand their purchasing behavior and tailor marketing campaigns accordingly.

Pricing strategy: analyze the effect of changes in pricing on product demand and to determine the optimal pricing strategy for different product lines.

Marketing optimization: evaluate the effectiveness of different marketing channels and optimize marketing spending.

Operations optimization: analyze the efficiency of operations, including the cost of goods sold, and optimize processes to increase profitability.

Customer satisfaction analysis: analyze customer satisfaction using the “rating” field and make improvements to the customer experience.

References

Dykes, B. (2019). Effective Data Storytelling: How to Drive Change with Data, Narrative and Visuals (1st edition). Wiley.