Faking Data
0 likes1,034 views
The document discusses strategies for creating realistic fake or "demo" data to demonstrate analytics and data visualization tools, including obfuscating real user data, generating random fake user profiles and activities, and programming the data to fluctuate realistically over time to mimic actual usage patterns. It also notes some data elements like URLs and titles that are difficult to fake convincingly.
![Obfuscate if you can. demoPayees = [“Joe's Widgets”, ” ACME Explosives”, “ Frank's Fish Fry”] transaction = { “ datetime”: “2011-07-13 18:52:22” “ payee”: “Churchkey”, “ ammount”: 145.23 } transaction.payee = demoPayees[rand(0,2)];](https://image.slidesharecdn.com/fakingdata-dcphp-july2011-110921090102-phpapp01/85/Faking-Data-9-320.jpg)
![Big Data GET /atapi/pub/shares/content?pubid=xa-4e1df9914f1a6d69 &dates=110706-110712&type=bm%2Cem%2Cfl%2Cif%2Cab &domain=www.example.com&range=0%2C20 [[" http://www.example.com.com/blog/2011/07/big-news ","Example Blog :: Big News",4990]]](https://image.slidesharecdn.com/fakingdata-dcphp-july2011-110921090102-phpapp01/85/Faking-Data-13-320.jpg)
![$shareMagnifier = $countrySharePercentage * $serviceSharePercentage * $browserSharePercentage / 100000; $combinationTarget = $domain['target'] / count($domain['urls']) / count($domain['referers']) / count($domain['keywords']); $dayMultiplier = (($daysOfData - $day) / 100) + 1; $combinationShares[-$day] = floor(rand(1,100) * $shareMagnifier * $dayMultiplier * ($combinationTarget / 50));](https://image.slidesharecdn.com/fakingdata-dcphp-july2011-110921090102-phpapp01/85/Faking-Data-22-320.jpg)
Faking Data
- 1. Faking Data (Building a Real-looking Demo Account) Philip Sharp @philipsharp Clearspring / AddThis DC PHP July 13, 2011
- 2. Screenshots are useful... Features/Tour
- 3. Help
- 4. Marketing
- 5. Screenshots are not a demo. Nor do they keep themselves up-to-date. (The same thing goes for videos.)
- 6. Neither is a free trial.
- 7. We collect data. People want to see how it looks up-front. (We're not all Google. GA doesn't need a demo.)
- 8. A good demo requires good data. (Chartbeat is lucky. But we can't all show Fred Wilson's data.)
- 9. Obfuscate if you can. demoPayees = [“Joe's Widgets”, ” ACME Explosives”, “ Frank's Fish Fry”] transaction = { “ datetime”: “2011-07-13 18:52:22” “ payee”: “Churchkey”, “ ammount”: 145.23 } transaction.payee = demoPayees[rand(0,2)];
- 10. Make your own if you can. { “ username”: “frodo” “ name”: “Frodo Baggins” rings: 1 }
- 11. Some data doesn't cooperate. URLs
- 12. Titles
- 13. Big Data GET /atapi/pub/shares/content?pubid=xa-4e1df9914f1a6d69 &dates=110706-110712&type=bm%2Cem%2Cfl%2Cif%2Cab &domain=www.example.com&range=0%2C20 [[" http://www.example.com.com/blog/2011/07/big-news ","Example Blog :: Big News",4990]]
- 14. Choose Your Inputs Services (top 50)
- 16. Ruleset: Domain
- 17. URLs
- 18. Keywords
- 19. Referers
- 20. 'demo-oz-emeraldcity' => array ( 'emeraldcity.gov.oz' => array ( 'target' => 5000 , 'urls' => array ( array ( '/' , 'The Emerald City' ), array ( '/tourism' , 'The Emerald City - Oz\'s Premiere Road-Trip Destination' ), array ( '/tourism/things-to-do' , 'The Emerald City - Things to Do' ), array ( '/tourism/things-to-do/balloon-rides' , 'The Emerald City - Things to Do - Balloon Rides' ), array ( '/tourism/things-to-do/horse-rides' , 'The Emerald City - Things to Do - Horse Rides' ), array ( '/mayor' , 'The Emerald City - Mayor\'s Office' ) ), 'referers' => array ( '' , 'google.oz' , 'bing.oz' , 'yahoo.oz' , 'search.oz' ), 'keywords' => array ( '' , 'oz' , 'vacations' , 'green' , 'emerald' , 'sunglasses' ) ), 'wizard.oz' => array ( 'target' => 3000 , 'urls' => array ( array ( '/' , 'The Wizard of Oz' ), array ( '/the-wizard' , 'The Wizard of Oz :: The Wizard' ), array ( '/seeking-audience' , 'The Wizard of Oz :: Seeking Audience With The Wizard' ), array ( '/contact-us' , 'The Wizard of Oz :: Contact Us' ) ), 'referers' => array ( '' ), 'keywords' => array ( '' ) ) ),
- 21. foreach
- 22. $shareMagnifier = $countrySharePercentage * $serviceSharePercentage * $browserSharePercentage / 100000; $combinationTarget = $domain['target'] / count($domain['urls']) / count($domain['referers']) / count($domain['keywords']); $dayMultiplier = (($daysOfData - $day) / 100) + 1; $combinationShares[-$day] = floor(rand(1,100) * $shareMagnifier * $dayMultiplier * ($combinationTarget / 50));
- 23. The Result
- 24. Using It Demo Accounts Demo Profiles
- 25. OOP to the rescue. /** * Hydra Package API demo wrapper * * @author philip * @requires Zend Framework */ class AT_Hydra_Package_Service_Demo extends AT_Hydra_Package_Service { … } $dataModel->setService(new AT_Hydra_Package_Service_Demo());
- 26. But wait, there's more. Browsers and Operating Systems
- 27. Last 24 Hours
- 28. This...
- 29. Looks like...
- 30. $radian = ($minute / 1440) * 2 * M_PI + (1.25 * M_PI); $sin = sin($radian) + 1.05; $shares = round(rand(1,100) * $sin * $shareMagnifier * $combinationTarget);
- 31. Dangers 0.00000001 * [Just about anything] is a really small number
Editor's Notes
- #2: Who am I. What we do.
- #3: People want to see the product. Our people want to show it off.
- #8: (save, edit, publish) manually, automatically
- #9: Point your site to a demo db.
- #10: Hard to obfuscate. Tedious to make manually.
- #11: cant store random stuff explicitly by-day
- #16: Date is rolling
- #18: passthrough where ok override where ok reformat data