R Project 1:
This project utilizes data provided in an analytics course, to measure actual return on ad spending.
An advertiser wants to know if online advertising will increase their store sales. A randomized experiment ("A/B test") was conducted using cities as the unit of measure by the publisher. Half the cities got the ads and half didn’t. Now we want to measure the return on ad spend (ROAS).
The A/B test lasted 30 days. The daily sales by city for 60 days prior to the test and for the 30 days during the test is provided as a data set. Additionally, the ad spend for each day (there was no advertising before the test started) is provided too. 10,000 cities were randomly assigned to test and control groups, half in each.
The questions to answer: Does advertising impact sales? If so can you quantify this effect? What does this effect mean for the client?
I have approached this problem as a multivariate linear regression, in which the ROAS is automatically derived from the results of the regressions. I have used RMarkdown and generated a PDF file from the R code:
More to come ... Stay tuned!