Data Analysis - Case Study - A Simulation
- Computer Science Department
Scenario:
Cecilia and Sam have just finished there ITSW-1307 class and are wondering how Data Analytics can provide value for their "Hot Dog" store. Sales have been good, but they both are aware that improvements can add to their bottom line.
In the Descriptive Analytics Approach, Sam and Cecilia analyze data aggregated across past weeks and months. Is the assumption that we sold "X" amount of hot dogs last week or last month sufficient evidence to make a decision on how many dogs and buns to buy for next week?
Clearly - this is better than Sam just saying "I have a feeling - next week sales will be great"
Together, Sam and Cecilia developed a descriptive analysis model.
Cecilia reminds Sam of the information in the ITSW-1307 course and suggests they incorporate additional data into their decision as to how much to buy next week. She suggests to Sam that they look at the pervious data as well as inject external data modules such as how the weather may effect sales AND is there a correlation between their Google search counts that may indicate an increase in business or upcoming traffic to the store.
She points out to Sam that when the weather is between 60 and 80 degrees, hot dog sales are at their peek. Sam agrees and they decide to include next weeks weather forecast into their decision. Together, they build a Predictive Analysis approach.
Sam and Cecilia have had a great amount of success in predicting future sales for optimizing the purchasing of hot dogs, buns, etc. They now are ready to take their analysis to the next step. They use their previous model and introduce: simulations, "what-if" scenarios, current traffic pattens, twitter mentions, and facebook likes using machine learning to their approach. They discover that there is a direct correlation if they are active on Twitter, and increase their following on facebook. Additionally, they learn that while sales increase dramatically during a "buy two - get one free" offer; hot dog sales drop dramatically the following week. They learn that the promotion is just a quick fix and based on the data.
Based on their simulations, machine learning, and computerized recommendation engines, they decide a loyalty point program will provide more sales and creates more return business.
Result - Business is exceptional and the they discover that their analytical model and analysis are a true winner! They franchise the "Hot Dog" store and two years later go public with their franchise in a stock IPO. They look back on their studies in the ITSW-1307 class and agreed that learning about data made the difference.