Updating Free eMetrics Ticket
Correcting the Rules
The rules should have said “up to 500 words,” not requiring 500 words.
- A chart illustrating something
- Up to 500 words about your Chart
- Up to 500 words why you would want to go
To submit your entry:
- Send an email to firstname.lastname@example.org
- With the subject “Submittal”
- INCLUDE YOUR CHART AND TEXT DOCS
- Required In The Email:
- Your Name
- Your Email
- Your Phone Number
- Your Physical Address
- Your Personal Website
- Your Twitter Account
- Your LinkedIn Profile
If you can get your point across in 20 words, more power to you for your brevity!
Submittals still go to email@example.com with the subject “Submittal” and questions to the same address with the subject “Question”.
Original contest post is available here.
Read on for the aweseome-ness of eMetrics, the recast UCI -MSNBC data and some charting tips!
Should I go to eMetrics?
If anyone is wondering about eMetrics, is it worth your time to submit: the simple answer is:
eMetrics San Francisco, along with San Jose on those alternate years, is the place to meet in carbon form all the measure professionals who you exchange tweets, Yahoo group messages and LinkedIn invites. That is on top of the incredible conference schedule which they have put together.
The Deep Data Diving track is particularly impressive, John Elder of Elder Research, and author of the seminal ‘Statistical Analysis & Data Mining Applications,’ along with other experienced practitioners get down to the hard core data analysis so many of us love.
Just getting started?
Take a look at the Web Analytics Fundamentals Track. The Free & Cheap Tools session is another session marked on my eMetrics conference matrix. When you are getting started out you probably can’t afford SAS, Tableau or other proprietary licensed software.
Leverage Free and Open Source Software solutions to drive your career forward!
Is Social Media your thing? Consider a [shameless plug] “Sentiment Analysis, Is It Ready For Prime Time?” panel discussion with Gary Angel of Semphonic and some other guy.
Everyone wants to know what their customers, prospects and competitors are saying about their product/service. Can this even be measured with the tools that are available today?
Something that I am working furiously to finish is recasting the MSNBC data into a more digestible format, most likely Excel files, with additional information as well as aggregated information for you to analyze. Follow me on Twitter where I will announce the release of the data later today.
Starting tomorrow we will study the recast UCI-MSNBC data for insights over the next several days to give you some tips on what you might look for.
Charts and Graph Advice
I have produced several charts and graphs that while not lacking in areas of improvement, should help people get started in terms of understanding what might be considered ‘good.’
Before we begin, if you have not already downloaded Eric Peterson’s free materials “Web Analytics Demystified” and “The Big Book of KPI’s” please do so now here. They will be a point of reference at times.
A few pointers from this chart to get you started:
- Descriptive Title
- A legend
- Clearly labeled axis
- Not too flashy – save the art project for someone who cares
- Use of the second Y axis is helpful in illustrating the relationship between the variables
Improvements to be made:
- Including every tenth day could have made for a better presentation
- Notations for the spikes – Holiday visitors?
- New, returning visitors did what?
- Low Click Depth, Medium Click Depth, and High Click Depth visitors did what?
Hope that helps a bit!
Pie Eating Contest
Please, please, please do not submit a pie chart.
- What am I supposed to do with this?
- What does this tell me?
“The Big Book of KPI’s” provides the basis for better charts, you did download it already right?
This combined chart starts to take a look at the trends of visits, first with a running seven day average and then with a daily percent change.
These are nice, but I feel that the you could do better. Its almost as if we needed to look outward, to an industry that has been charting data for a really long time.
NYSE and S&P Historical Data
Suppose you have a two stocks, StockA valued at $10 and StockB valued at $100. The next week StockA is valued at $20 and StockB is stuck at $105. StockA increased 100% but StockB’s starting location will make that gain appear insignificant.
If you were to compare these two variables you may end up with the top of this chart showing historical Dow Jones and S&P data, the Dow dwarfs the S&P but we are potentially overvaluing the fluctuations in the Dow index.
Common in finance and economics is semi-log plotting, for the second half of the chart the same indices are plotted in a semi-log scale. That is where the dependent variables, along the x-axis, are transformed logarithmically while the independent variables, along the y-axis, are left alone.
Log Scale in Excel
To perform a log transformation in Excel the code would be:
=log([cell containing data])
The final chart showing visits and pageviews in a semi-log scale allows us to evaluate changes in visits and pageviews on the same scale, removing the difference built into the arithmetic scale.
If you can’t tell by now, I really like log scale plots. [hint]Really like them. [/hint]
Don’t forget that the recast UCI-MSNBC data sets will be released later today, follow me on Twitter to make sure you catch it while its hot.
Tomorrow we will take a look at those data sets, what we might be able to glean from them and some cool charts using Excel.
Questions, ideas? Email me at firstname.lastname@example.org.