Overview

Sprint lead: margaret
Sprinters: sdwilsh

Description

A statistical analysis of the user's history, average tab load, etc. Like Google Zeitgeist, but based on their Places database.

• Dietrich has an add-on that does some of this already. screenshot

Goals / Use Cases

• Reflective personalization: Showing users the broader patterns in their interactions with the web, thereby creating a tighter personal bond with it.
• Debugging: testers can use this page to gauge their web usage, and the amount of metadata generated by it, for feedback on bugs and feature usage.

Non Goals

• Drill-down or query-ability of the data past the basic display on the page
• Remote aggregation and analysis of this data

Prototype Development

Here is a list of brainstorming ideas for the about:me page. The latest prototype is available as an add-on, which you can download here: https://addons.mozilla.org/en-US/firefox/addon/13681.

Fun

Activity stats

• Hourly trends [DONE]
• Daily/monthly trends
• Top sites visited [DONE]
• Total usage over time
• Average time on a page
• How you get to pages (bookmark, link, etc)
• Relationships between different pages (parent, etc.)
• Novelty of Browsing (% first visits)
• Recent activity
• Sites visited multiple times but not recently
• Trends over total history

Tab stats

• Lifetime
• Frequency visited
• Average number of tabs open
• Information about current tabs

Bookmark stats

• Top used bookmarks
• Tag cloud for bookmark tags
• Hourly/daily/monthly trends

Extensions

• Installed add-ons
• Usage patterns (how much do we know about this?)

Technical

General

• IP address
• Physical location
• Internet connection
• Network usage
• Plugins
• Cookies
• Current profile name

Memory usage

• Usage for different tabs/sites

Download stats

• Number of downloads [DONE]
• Daily trends [DONE]
• Types of downloads [DONE]

Implementations of Similar Things

• Google Zeitgeist
• Google Reader Stats
• Google Web Activity Stats
• Flickr Stats
• Chrome about:stats

Feedback

• Include thumbnails for website entries
• Count unique visits per day instead of total visits
• Distinguish between sites with significant subdomains (e.g. google.com and google.com/reader)
• Remove default pages from page (e.g. default google start page)
• Analyze trends within specific domains

Design

Original design ideas for some potential sections and datapoints.

History & Bookmarks

• number of history visits

SELECT COUNT(1)
FROM moz_historyvisits
WHERE visit_type NOT IN (0, 4)

• number of unique URLs

SELECT COUNT(1)
FROM moz_places
WHERE hidden <> 1

• fav sites (top 5-10, total visits, avg visits per day)

SELECT *
FROM moz_places
ORDER BY visit_count
LIMIT 10

• avg visits per day

• day of week stats (eg: "sat broader, sun has more revisits") - would need total visits (sunday is 0)

SELECT COUNT(1)
FROM moz_historyvisits
WHERE visit_type NOT IN (0, 4)
GROUP BY strftime('%w', ROUND(visit_date / 1000000))

• number of unique URLs for each day of the week (sunday is 0)

SELECT COUNT(*) as num_visits, 
strftime('%w', visit_date/1000000, 'unixepoch', 'localtime') as day 
FROM moz_historyvisits v LEFT JOIN moz_places h ON v.place_id = h.id 
WHERE v.visit_type NOT IN (0, 4) 
GROUP BY day ORDER BY day 

• graph of history volume over time (visits per week since start of history accumulation)
• total browsing uptime (via history, maybe goes in App)
•  % of the Web visited (out of 1,000,000,000,000 pages)
• url and date of oldest visit
• top URLs most selected from the awesomebar
• number of bookmarks

SELECT COUNT(1)
FROM moz_bookmarks
WHERE parent <> :tag_root
AND parent <> :places_root

• url and date of oldest bookmark (oldest by what?)
• graph of bookmark activity over time
• number of tags

SELECT COUNT(1)
FROM moz_bookmarks
WHERE parent = :tag_root

• revisitation
  • average or median visits per URL
  • percentage of visits that are bookmarked or tagged pages

Downloads

• total bytes downloaded

SELECT SUM(currBytes)
FROM moz_downloads

• number of downloads

SELECT COUNT(1)
FROM moz_downloads

• date of first download

SELECT startTime
FROM moz_downloads
ORDER BY id ASC
LIMIT 1

• average download size

SELECT AVG(maxBytes)
FROM moz_downloads
WHERE maxBytes <> -1

• graph of download activity over time

• average download time

Other

• avg num open tabs
• current tab load
• average tab load since XXX
• avg startup time
• number of cookies
• entries and total bytes of cache (or link to about:cache)
• number of extensions installed
• number of plugins installed (or link to about:plugins)
• IP address
• user's physical location
• search efficacy (original blog improved sql )
• number of saved passwords
• degree of variation between saved passwords

Bugs

bug 480154