My Most Challenging Analysis - Issue 176
Reflections and learnings from my most difficult analysis.
Welcome to the Data Analysis Journal, a weekly newsletter about data science and analytics.
What was your most difficult analysis?
It used to be my favorite ice-breaker to ask analysts to learn what they struggle with the most.
Some of the answers I remember:
Building graphics in R.
Developing a multiple linear regression for financial forecast in Excel.
Working with raw unstructured and non-linear data for analysis.
Running A/B tests with very low-traffic sites.
Applying the Random Forest model for regression tasks.
Mapping or merging data sources that don’t share any common values or attributes.
Loading and parsing XML files in a structured table in db.
Forecasting LTV for businesses that offer 100 types of trials.
In most of these responses, the challenge wasn’t related to the complex math behind calculations or understanding the sequence of steps. Instead, the challenge was balancing one’s time for a project with underestimated deadlines.
Today, I wanted to share two of the most challenging projects I’ve had to work on. First, I briefly touch on the price change analysis, then focus on measuring a new feature impact from an A/B test affected by network effects. Alongside, I will share my learnings and reflections on the process, analysis, and mistakes we made - one of which was underestimating bias.
Price increase analysis
Keep reading with a 7-day free trial
Subscribe to Data Analysis Journal to keep reading this post and get 7 days of free access to the full post archives.