Data Spreadsheet finally updated; GIS working

Mid-February, ir looked like I was ready to use my data on my project on Theodor Tantzen and the German Democrats. Right On examination, I founds that some of my data was incomplete, particularly for the districts around the city of Oldenburg. I also compiled a list of all my missing results and diligently went back through the copies that i had made of polling place results. I was really assisted in this by PDFs of different Oldenburg newspapers published online by the Landesbibliothek Oldenburg (thanks!) I have described this process and some of the changes that I have made to my map and data on this site under "Current Projects".

One new matter. QGIS has been running quite slowly. I have learned that I can make it rune more smoothly if I correct the projects geometries. The checked geometries feature indicated that I had nearly 12,000 incorrect geometries. When I used the "fix" function, one error would be "fixed" while more would appear. I have begun to corrected this problem by systematically deleing individual polygons and creating new ones using a different method. What little I have been able to do, indicates that this will speed thing up considerably. I hgave also decided to experiment with new formats, i.e, SpatiaLite, PostGIS, as time permits.

One problem solved

I scrolled through the votes by polling place for the problematic second election of 1924. I found that when I ordered rows by polling place numbers and look at the handful of major cities, all of their votes were recorded on a row lower than they should have been. The errant entry was the village of Blandorf, quite high on the list. When I adjusted the row one up, all of the cities lined up. Searching towards the bottom of the data base, I found an entry for Blandorf with all of the correct voting information but somehow given an Oldenburg polling place number. Problem identified and solved.

While I had the early archived list of votes results, I took the opportunity to correct another shortcoming to my Project_Master. When I created the version that I am now using, I did not include the Reichstag elections for 1898 and 1907. Those votes have now been entered and my next step will be to calculate the percentage of votes cast in each place for the competing parties. Once that is finished, I will include the several special elections held in the Imperial period (when a deputy either died or resigned). I had not added the later in earlier version as they seemed to tax the permissible columns in xls. As I have added material lately, this does not seem to be a problem. I wonder what is different? If space is available, it might be interesting to find and add results for the presidential election of 1925 or the various plebiscites that were held during the Weimar era.

OOPS

At the new year, I thought that I had resuscitated my GIS. Not so. The shapefile still needed work around the edges. As in any project, its those outstanding edges that were the most difficult.

When that task was finally complete and I tried to populate it with my data file, I noticed that many polling places were showing no results. This could not be the case, as I had notes on the handful of polling places for which I did not have data. A meticulous review of my Project_Master file indicated that perhaps 1/8 of my data was missing. How? Why? Did I lose data in the conversion for xls to cvs (required by QGIS?

Fortunately, after much searching, I was able to find an uncorrupted backup file. I made a spreadsheet of the missing polling place data and am presently integrating the "lost" data into my Project_Master file. Doing so, I uncovered another problem. The results for the second national election in 1924 were jumbled. I will have to diagnose that problem before I continue.

The creation of the original GIS map took +3000 hours; creating the data file (not counting the time collecting data) took a further +2000. I am comforted that all of this work was not lost!