16. Stay focused!
The data trinity by Kaushik
What is happening?
- Key metrics
- behavior analysis
- segmentation
Why is it happening?
- Understand user experience
- customer feedback
So what?
- Efficacy
- conversion
- revenue
- world peace?
21. Data tips and pitfalls:
• All data is interesting, only some is actionable!
• Focus on important outcomes
• Start quick and dirty with what is available
• Take a mixed methods approach
• If you torture the data long enough it will say
anything! i.e. data lies
• Leave room for creative freedom
22. Tips for creating a data driven culture:
• Regular reporting
• Set goals around development, marketing, etc
• Dashboards (ala Leftronics, Geckoboard)
• Secure a spot for data in decision-making
• Share graphs via social media and paper!
• Give the data person access across the team
• Leave time for data exploration…it’s fun!
23. Sites and resources
• Book: Data Analysis Using SQL and Excel
http://www.amazon.com/Data-Analysis-Using-SQL-Excel/dp/0470099518
• Top Gifts for Data Scientists
http://www.forbes.com/sites/ciocentral/2011/12/22/top-holiday-gifts-for-data-scientists/
• Data Jujitsu (for data products)
http://strataconf.com/strata2012/public/schedule/detail/23092
• Free, very basic SQL tutorial
http://www.sqlcourse.com/
• Online programming courses
http://www.codeschool.com/
• Data conferences and events
http://www.kdnuggets.com/meetings/
The MVP (minimum viable product), lean startup approach to rapidly iterating technology also applies to how we handle data. Do quick tests, find available data, do adhoc analyses, guess and check, focus on quick wins and low hanging fruit 80% of the time. When you’ve hit on something really important, that’s time to dig deeper.
One area of stats that is very popular in the data science community is predictive modeling: using available data to predict the probability of future outcomes. If you are interested in this work there is a whole, week long conference about predictive analytics every year, and a huge industry springing up around this field.
Data products don’t have to directly make you money! They can help attract people to your site, like this super cool social graph by linkedin.
Data visualizations are another big area of data science. They help us represent huge data sets in ways that make sense for internal analysis and also external communication.
Data visualizations are art! There’s no doubt about that. And truly Information is Beautiful (Google it!). So whether your data passion is in art or in engineering, there’s room for you in the data world.
iPhone users beware, ahaha! Even super simple visuals can create a top blog, like OkTrends, or get you a first page spot on the WSJ.