@Lewis Some more info, but out of the order in which you asked:
>> session reconstruction e.g. What does an end-to-end browning session >> look like? Browsing session reconstruction defines a good amount of the work we’re doing now and for the remainder of the year on a number of projects. We’re actually already working on this now, and it will be the basis of a ton more work that will start mid-late June. What our solution looks like is UserALE.js packaged in a FireFox browser plugin. The plugin drops UserALE.js into the client without having to touch source code. So, you can get all the data you’d expect from UserALE.js, but be able to capture data from sites that UserALE.js script tags haven’t been embedded in (the source code). The use case is that orgs and users can just deploy a browser that captures userALE logs with disc images—bam. Chris Muto (new contributor) built a prototype. @fordar is picking that project up and resolving some of the major/obvious issues like mixed media/CORS issues. You’ll see this code in the UserALE.js repo probably next week when @fordar get’s his bearings. >> what type of client e.g. Firefox, safari, etc Plugin will be FireFox to start. But FireFox and Chrome are playing nice in plugin portability… Once the plugin is in the .js repo, feel free to reach out or combine forces with @fordar if you know of someone that needs this very soon. >> heatmap generation representing user activity overlaid over Webpage >> * dot plotting similar to the above First... what Rob said Second, I actually have some tickets to move over to Apache Jira about investigations of open source code for grabbing screen caps from browsers. If we can coregister those screens with cursor x,y, then heat maps should be a snap (if you can deal with pushing around and storing those screen caps). The browser plugin logging solution will also make this easier, building off the plugin should give us more data available to the client. Next week, I’ll get my investigation tix moved over (going on short vacation tomorrow—back next week). Unlikely we’ll have a ton of bandwidth (as a team) to work on heat maps specifically in the next month or two. However, later this year we’ll be focusing very hard on visualization and TAP. Something like heat maps are a big part of that, but I can’t talk about that with specifics yet. @lmariano is very interested in this, but she can’t talk specifics yet either. Again, comprehensive browsing session capture is a major theme for us this year. A number of us will be under contract to do some of this work and it will all be Apache. If you know of other interested parties, let’s talk offline so that we can coordinate sooner than later. Things will start percolating up to our roadmap pretty soon. >> * ability to see where people visiting your webpage/webapp came from e.g >> where did the client come from? No immediate plans to do this. No current capability to capture origin/destination data, but likely something that could be deployable through the script tag. We should have some wish list tix on this. The browser plugin deployment use case will drive a lot of upgrades to UserALE.js that will push this along. Of course, we’ll get this data from the client easy, but the “visiting your website” deployment is different than the “websites you visit” use case the browser plugin supports. I hope that helps, too. > On Apr 27, 2017, at 10:48 PM, Rob Foley <rob.foley...@gmail.com> wrote: > > Hi Lewis, > > While I haven't taken to plotting it out myself, I think the data we > generate would be sufficient to plot out a heatmap on a page, since many of > the logs contain X,Y coordinate pairs. Using a library like heatmap.js or > simpleheat you could probably get a decent overlay for a page that way. Dot > plotting would be a similar answer with something like D3. Seeing where > someone came from may be a bit tricky, since that'd rely on the referer > (sic) header being set, which some clients may ignore. If you only wanted > to track between pages on a particular domain, that may be a bit easier. > > User Agent is currently picked up by our logstash logging endpoint, though > any logging server could record it. For a better idea of what the client > looked like (client fingerprinting), it may be better to perform various > feature detection heuristics. Reconstructing a browser session is a bit > ambiguous, since it may require even more fine grained recording, though > it'd probably make a fairly good estimation. > > Hope that helps. > > Rob > > On Thu, Apr 27, 2017 at 9:26 PM, lewis john mcgibbney <lewi...@apache.org> > wrote: > >> Hi Folks, >> Can someone explain the capabilities offered by senssoft for the following >> scenarios >> * heatmap generation representing user activity overlaid over Webpage >> * dot plotting similar to the above >> * ability to see where people visiting your webpage/webapp came from e.g >> where did the client come from? >> * what type of client e.g. Firefox, safari, etc >> * session reconstruction e.g. What does an end-to-end browning session >> look like? >> >> I have plenty more questions but lets start with the above. >> Thanks >> Lewis >> -- >> http://home.apache.org/~lewismc/ >> @hectorMcSpector >> http://www.linkedin.com/in/lmcgibbney >>
