Commons:British Library/Mechanical Curator collection/georeferencing campaign/anomalies

There are various anomalies between numbers based on Flickr searches, on the main Google Fusion Table, and on the BL maps page.

Numbers (snapshot)

As of 01:00 BST, 15 April:

• The main fusion table is saying 12205 done, from a total of 49912
• The Commons progress page is saying 12209 maps done, and (implictly) 38004 to go, from a total of 50213. (Possibly a very little double-counting, due to a few books appearing on more than one index wikipage).
• The geoparse table has 15341 returns (includes about 3137 from phase 1 -> abt 12204 from phase 2)
• Flickr has 12133 maps tagged hasgeoref + georefphase2, 36,282 tagged togeoref + georefphase2, from 49974 maps tagged "georef_phase2" in all (which includes 1565 tagged "georef_phase2" with neither "has georef" nor "to georef".
• The progress bar at http://www.bl.uk/maps/ is reporting 13106 maps georeferenced and 37107 maps to go, making a total of 50213 in this phase.

Some considerations

• A check in April 2015 found 276 images marked "to georef", but the georeferencer link gave a 404 (example -> 404). This gets close to bridging the gap between 49912 in the Fusion Table, versus the supposed 50213.
• Maps that have been split, or are parts of a world map, have not been tagged "to georef" -- eg: UK examples (83), more examples. These were not supposed to have been included in the georeferencing run, but in fact were.
• Flickr tag counts can be unreliable, especially for complicated searches: eg 37 supposed maps "to georef" but not from "georef phase 2" -- but in fact they are.

Previously missed maps

Normally only maps tagged since the most recently processed tagged maps are considered. However, following a re-upload of the Fusion Table, the whole lot were considered for tagging and for geo-parsing.

This gave:

• 133 more maps tagged "has georef"
• Geoparse entries went up from 14923 to 15341 (+ 418)