The Roadblock Database

Please note that the database itself will not be made public for the time being!

Creating the database

The first job of the roadblock project was to create and populate a new database with the entire 1941 survey. This was no easy task; the survey comprised 473 roadblocks, each of which had 35 fields of information associated with it, making a total of 16,555 values (including null values) to be accurately entered.

I decided to create a bespoke database rather than just import the data into the Documentary Evidence database in order to be able to record the individual features of each block, such as number of cylinders, buoys, pimples etc which cannot be done in the exisiting database.

The data forms included with the report provided an obvious basic structure with field names, but a few extra fields were needed to cater for non-standard obstacles such as extra-large cylinders.

The data entry process took about 10 days to complete and was done using a ruler on each page of the report to keep my vision in a straight line. Problems arose with unclear entries or where my photograph of the original page was not of best quality; a second look at the original document was needed to resolve these issues. The main problem was an error on the part of the compiler, who inadvertantly swapped two columns round on a couple of pages. Fortunately I saw this error before I had transcribed too many entries in the wrong fields in the database and was able to backtrack.

Data verification

Once this was all done, I then had to verify and correct the data entered. Fortunately, the report had several statistical tables of totals of materials by sector, allowing me to find data input errors reasonably quickly. In all, I uncovered about 30 mistakes on my part - not bad out of over 16,000 values!

The other problem was where the report's compiler had himself made errors; going back over the raw data, I found a few oversights, but some mistakes could not be accounted for. The table below shows the full number-crunch; a crossed-out value is where I have located an original error that cannot be resolved.

	Block total	Cylinders	Sockets	Bent Rails	Straight Rails	Bent RSJs	Pimples	Buoys
A Sector
Current Roadblocks	40	361	143 148	22	171	117	269	574 564
Redundant Blocks	21	103 91	12	0	12	0	15	161
Sector Totals	61	464 452	155 160	22	29	117	284	735 725
B Sector
Current Roadblocks	72	1034	654	200	100	170	239	380
Redundant Blocks	55	636	314	89	165	48	343	673
Sector Totals	127	1670	968	289	265	218	582	1053
C Sector
Current Roadblocks	84	750	906	102	2012	504	521 539	1443 1451
Redundant Blocks	44	267	275	81	173 92	54	117	447
Sector Totals	128	1017	1181	183	374 293	554 558	638 656	1890 1898
North Sector
Current Roadblocks	88	1454	62	17 16	273	0	127	570
Redundant Blocks	69	231	100	45	134 147	0	17 23	554 572
Sector Totals	157	1685	162	62 61	161 174	0	127 150	1126 1142
Total	473	4836 4824	2466 2471	556 555	829 761	889 893	1648 1672	4804 4818

1. Seemingly includes 6 straight RSJs.
2. Seemingly includes 8 long straight rails.
3. Seemingly includes 5 long straight rails but not 7 straight RSJs.

A fascinating dataset, but only part of the project. Using the roadblock database, I intend to visit as many locations as possible to see what evidence still remains. The database itself will not be made public for the time being.