Update: see follow up issue https://github.com/piwik/piwik/issues/5246
Most popular pages/keywords/etc. are popular every day, and it is hard to know what changed since the day before. What would be interesting would be to visualize:
This could help answer questions such as:
This report could be available as a widget showing the top 5 movers and shakers overall (increased keywords, pages, etc.).
We could maybe improve at least the “basic” view showing only visits.
In this view we could for example display in a new column on the right, a green or red % increase, when there is a a major change (threshold set to +/- 10% for example). Otherwise leave blank. See also #291
We could put threshold of minimum visits impacted = 20
minimum % increase or decrease = 2% for example?
So it would not display minor changes that are not interesting.
They could be displayed on top of the report
(green sign) India (50%) and Scotland (33%)
(red sign) France (-33%) and Germany (-44%)
| Country | Visits |
| India | 176 |
| Scotland| 54 |
| France | 23 |
| Germany | 23 |
The code would be optimized so that it returns in less than 2 seconds when calling 8 different APIs so comparing 16 datatables.
The algorithm will typically extract:
New widget to show this great insights:
see GA hack on http://www.juiceanalytics.com/writing/keyword-trends-google-analytics-greasemonkey/
To produce significant movers and shakers, the algorithm could maybe select "current month" dataset and compare it to yesterday if period=day, or to last week if period is week, for example to find out the items with an average very different from the number "this week" ?
See also relevant article: http://www.wired.com/magazine/2011/06/ff_feedbackloop/all/1
ALso interesting: http://www.juiceanalytics.com/writing/keyword-trends-google-analytics-greasemonkey/
display_limit: This controls the maximum entries that each table will contain. This may be useful for large, sprawling sites.
growth_tolerance: This is the percentage growth parameter. Changing it to .10, for example, will catch everything that has grown by 10%, as opposed to the default 50% and 20%, respectively.
minimum_number_elements: This is a significance benchmark that can be used to limit what is displayed upon the screen. By default, only keywords with at least 10 elements are displayed upon the screen. Referring Sites does not have a minimum by default, but one can be set if desired.
limit: Limit is more of an internal parameter that determines how many entries should be downloaded from Google in order to get the results that are visible here on the page. Lower the limit to increase speed. If the limit is set to a very high number, you will get the largest result set, but you will have to sit around for a while for the results to load. Since the results are downloaded ordered by volume, raising the limit from the default numbers will not actually give more significant results. You will simply get more of the smaller results, such as keywords with only 1 hit.
look_back: This is a very important parameter. The script uses the date displayed upon your Google Analytics page to determine the full range that you want to consider in your results, but look_back determines how many of those days are used for the significance test. So, say the range you have displayed in Google is March 23 April 22 and your look_back is 7 days. The script will compare the average referrals for a given keyword from April 16-22 to the average from March 23-April 15, and will return the keyword only if the recent average is 20% higher than the rest of the time period. Thus, if you want to increase the total range of the data, change the dates on the actual webpage. Change look_back only if you want to change the period of significance.
See also #3032 Trends reports
In general, this feature has much potential. Maybe other developers have better ideas to surface interesting data from the raw reports. We could write a simple plugin that would use the raw data to show insights. Other developers could then implement their own algorithm very easily by using the piwik APIs. We could even do a blog post and invite developers in a contest to write algorithms?
We could have a system that gives out several data insights profiles:
See also KPIs E-commerce : quels indicateurs pour votre tableau de bord for inspirations on good metrics that matter (french)
see also nice ideas in http://www.kaushik.net/avinash/six-visual-digital-marketing-analytics-strategic-solutions/
Consolidating milestones FTW
In 3055a65: refs #5477 started to work on displaying Insights. Added a widget for the dashboard and a new Insights visualization. More to come... still early version so expect a lot of things to change but I am still happy about feedback
In 3f2916b: Enable Insights plugin for all beta testers in the coming weeks. Please send us your feedback about this new Magic Feature! refs #5477
In b062cb7: refs #5477 added tests and fixed some bugs
In 7714334: refs #5477 do not call Insights API in integration tests
In 2a60ff3: refs #5477 fix tests
In 8110e9d: refs #5477 added possibility to place a widget multiple times (in case you want to see the same wdget with different visualizations and/or different parameters like different filters to make it easier comparable), removed some controls and save parameter changes in dashboard
In 727da1b: refs #5477 some bugfixes and started to work on insights widget
In 4de7581: refs #5477 fix tests
Many thanks for your work on this!!
two tiny comments for looking inside on the table sites:
using green and red as font colour for - and + values is great.
Maybe one could visualized the differences even more by using some kind of bars directly in the table.
Put an rough concept as attachment.
In 856a3e6: refs #5477 reducing memory usage and improving performance
In cbe88c1: refs #5477 destroy all tables but the insights tables
In 6e022a7: refs #5477 also delete the last table
In 62939c1: refs #5477 do not flatten for now, brings massive performance improvement 10min > 1 sec as Referrers_getKeywords takes forever and also leads in memory exhausted when report is flattened
In 6034c0e: refs #5477 fix tests
In 23186d7: refs #5477 make sure text ellipsis works and label is always only one line. moved test fixture to insights plugin as demonstration for other plugin developers as it is very insights related anyway
In 8601c09: refs #5477 some more adjustments to the widgets and added a less file
"broken after update to 2.1 final"
because disabling the inside plugin fixes it...
In c25f396: refs #5477 started to work on movers and shakers and lots of minor improvements
In a9e8651: refs #5477 some refactorings to have better testable code, loads of tests and bugfixes
In 4bb969c: refs #5477 added more tests and some bugfixes
In 862e532: refs #5477 removed min impact control as it might be not clear what it is anyway and user can change the limit filter if he wants to see only entries with very high impact
In 593dbc8: refs #5477 some UI tweaks
In ad9ab5d: refs #5477 more tweaks to insights and movers and shakers and more tests
In 4165176: refs #5477 some code improvements and tests, also some adjustments for sites having not many visits
In fc6fe7f: refs #5477 make sure always a change of at least 2 visits is required. Added icon license information, will move the information later
In aa40466: refs #5477 fix syntax error
In d210180: refs #5477 ui tweaks and translations
In 81c3786: refs #5477 moving translations to general en.json
In 78e9614: refs #5477 average filter is no longer needed
In 3710bb4: refs #5477 some more code tweaks and documentation
In a4f6fc5: refs #5477 moved legalnotice info for insights icon
In e96399a: refs #5477 added missing translations
In 2e1cb7b: refs #5477 make sure all column in overview widget have same width
In 9af9ae3: Refs #5477 New good looking icon for data insights for Morpheus theme!
Movers and shakers.png
Insights view of a report.png
First version of this feature is now available!
Leave us your feedback, ideas, suggestions in this ticket: #5246 Data Insights, Movers and shakers: improvements and suggestions