3. A G E N D A
01 A Data Science Challenge
02 Foundations
03 Data
04 Results
05 What Can We Conclude?
4. Where - in what U.S cities - should
a company conduct targeted, localized
marketing?
The Challenge
5. 1. Identify cities where our company is doing well.
2. Gather location-specific data.
3. Determine drivers of performance.
4. Use these drivers to inform business strategy.
And Approach...
15. Where are customers located?
Outcome Variables:
1. density: customers per relevant local population
2. growth: rate of local increase in customers
Local Marketing Metrics
16. What are the local sources of customers?
Local Marketing Metrics
17.
18. What is the local performance of
marketing?
Local Marketing Metrics
21. ● Young or old population?
● Internet use?
● Educational attainment?
● Employment by industry? Tech? Restaurants? Arts?
● Local business characteristics?
● Macro-econ factors: GDP growth? Wealth? Income?
Local Marketing Metrics
23. What did we find?
● % of people 30 to 44 not younger
● % of people with graduate degrees
● number of food carts!
24. New York
Atlanta + Houston + Dallas
Philadelphia + Baltimore
West Palm Beach + Miami
City Clusters
25. 1) Focus on locations that our analysis predicts should do even
better than they have so far:
We’re calling this “slack”.
“slack” x “population” = “opportunity”
2) Focus on locations that are outperforming our expectations
AND still growing at a rapid rate.
3) (Ideally) pick cities from different clusters when we compare
them across the dimensions identified in the model.
How did we use these results?