We're working along very similar lines, [rest omitted upon request, converting to robot mode]
Example (click link) of converting public data locked up in PDF to spreadsheet form and doing it collaboratively.
Tools used: [PDFtk4all][2] to split v.large heterogeneous PDFs into homogeneous sections, [Tabula][3] to convert PDFs.
[Recent bug-fix in Tabula][4] that makes the transcription process far more accurate, would advise that version for anyone dealing with non-English letters.
Working example (click link) to combine data in excel/csv with its geo-location data that is residing in a separate file and render them together on a map using leaflet.js.
Problem this addresses: Difficulty to add data that one may want rendered on a map to an existing geojson/kml format file. With solution, much more convenient to have the location data, and information to render, in separate files with latter being a simple CSV that is easy to edit. Potential for live updating of data by a team on shared google spreadsheet, publishing upon changes in CSV form, directly hooking up with map on website, hence eliminating workload.
Request to question-poster for contact for chance of further collaboration and shared exploration. In absence of any direct communication method on opendata.stackexchange
(kindly guide if so), sharing email address for contacting: nikhil.js [at] gmail.com
Removed 3 links since not allowed to post more than 2 links. (Not understood why points earned on other stackexchange platforms not carrying forward to here)
Specific question posted by asker on sharing best practices and examples, answered in this answer. Hoping moderators find satisfactory and don't take umbrage to effort to connect (which might only increase sum total of open data in the world if permitted) or misunderstand as CV-pushing (seriously, some people just don't need that where they are in their life right now).