librariesio / libraries.io

:books: The Open Source Discovery Service
https://libraries.io
GNU Affero General Public License v3.0
1.11k stars 202 forks source link

Gather moar data to support use case production #1183

Closed BenJam closed 7 years ago

BenJam commented 7 years ago

Following from #1127 and #1177 it would be good to gather a lot more data from analytics to support/dispute some of the assumptions we make of our users. This will involve

BenJam commented 7 years ago

right, now that #1184 is done... we can move on.

Search

On a sample size of 301k sessions 268k came straight from Google search into a project page, mostly into a /github url (131k)

what are the most popular search terms?

Looks like most search terms are for specific projects rather than keywords, top three being casperjs, phantomjs and angular

what are the results for those search terms?

The specific examples above result in the npm main project.

does a user click on any of those search terms?

32k performed a search, 21k of which dropped off at that point. The rest clicked through to a project, mostly on npm.

The only real conclusion I can draw here is that search is currently outsourced to Google and that it could be improved massively, thought there's a cost/benefit analysis needed there.

Package manager support

what are the most popular package managers?

NPM, Bower, NuGet

what package managers that we don't support should we add next?

Needs addressing in #1176

Maintainer functions

Increasing contributions

I'll keep adding to this ticket, feel free to add anything that might be insightful.

BenJam commented 7 years ago

@andrew measuring scroll depth on a lot of current pages would be useful using https://github.com/robflaherty/jquery-scrolldepth or https://github.com/leighmcculloch/gascrolldepth.js. Any aversion to my adding these?

andrew commented 7 years ago

No aversions here!