ICC Freight Commodity Statistics for the Boston & Maine Railroad

Index

• Background
• Collected Statistics - Major Variations in Commodity Codes
• Commodity Code Notes, including minor commodity variations
• Changes in Data Collected for Each Commodity - National/Regional Statistics
• Changes in Data Collected for Each Commodity - Individual Railroad Statistics
• Process
• Database Data
• Database Design
• Common Abbreviations
• Copyrights of Original Data
• History of this web page

Background

Starting in the year after it was created, the Interstate Commerce Commission collected statistics on the railroads in the United States.

Between 1888 and 1976 (possibly later, I just haven't researched beyond 1976) the ICC collected and published statistics on the railroads in the United States. Freight commodity statistics for individual railroads, including the Boston & Maine RR, was reported between 1917 and 1963. The numbers are broken down by "commodities" (a definition that changed with time - see below), and each commodity was broken down further by "direction" of traffic (again, see below).

Note after 1963 the ICC continued collecting the statistics, but they only published "regional" data, not data of individual railroad numbers.

If you wish to see the raw data, libraries in most major cities probably have the reports, typically in the "main" branch (or the branch where they store government records). Ask for:

1. United States. Interstate Commerce Commission. Bureau of Statistics. Annual report on the statistics of railways in the United States to the Interstate Commerce Commission
   Years published: The Buffalo and Erie County Public Library (New York) had 1888 - 1953. I do not know whether or not this particular report was published after 1953.
   Individual railroad commodity statistics are published in this report starting in 1917, other individual railroad statistics started earlier.
2. United States. Interstate Commerce Commission. Bureau of Statistics. Freight Commodity Statistics, Class I Steam Railways in the United States for the Year ended December 31, [year]
   (years published: 1924 - 1957)
3. United States. Interstate Commerce Commission. Bureau of Statistics. Freight Commodity Statistics, Class I Railroads in the United States for the Year ended December 31, [year]
   (years published: 1958 - 1976)

Collected Statistics - Major Variations in Commodity Codes

Over time, the data the ICC collected varied, both the commodity classifications and how each commodity was broken down. There are four major variations in commodity codes during the period when the reports listed railroads individually:

• 1917 - 1919 (although the same commodity codes were used since at least 1914, individual railroad data was not reported until 1917)
• 1920 - 1927
• 1928 - 1946
• 1947 - 1963

Starting in 1963 they changed the commodity breakdown again, and started using commodities codes compatible with the SIC - "Standard Industrial Classifications" commodity codes. These are NOT shown in the table below, as the data was not reported for individual railroads.

In the table below, I attempt to match up commodities as makes sense to me. Let me know if you think I am in error. Please note that I realize some matchings are not "exact", I tried to match as closely as possible given my [somewhat limited] knowledge (someone really needs to review the ICC documentation to see if they note how to handle the commodity changes).
Numbers in parentheses "(...)" are notes - see the list below the table.
Blue Numbers in brackets ("[1000]") are _my_ unique commodity codes (i.e. "Database Code") (something I added to make database queries a little easier - see note (8)).

Commodity Code Changes

Commodity Code Notes, including minor commodity variations

1. I only researched back as far as 1914. The dates for this scheme of classification may go back earlier than 1914, and there may be other, earlier classifications.

2. Depending on the table and year, these "Commodity Codes" may or may not have been listed in the reports.

3. Depending on the table and year, these "Commodity Codes" may or may not have been listed in the reports.
   The codes for "Total" for each group were never specified.

4. From the footnotes for this commodity:

   Includes automobiles, motor cars, autotrucks and parts (except tires and chains); chassis, tonneaus, and wheels

5. Note that in the printed reports from 1939-1946, this was included in the ICC data category 701 data.

6. In 1942 and 1943, the data for certain commodities were NOT reported individually.
   Per the "Introductory Statement" in those years' reports, these data WERE included in the totals "Grand Total, Carload Traffic" (commodity 850) and "Grand Total, Carload and L.C.L. Traffic" (commodity 860):

   ICC Code	Description	Database Code
   330	Copper ores and concentrates	3200
   331	Lead ores and concentrates	3201
   332	Zinc ores and concentrates	3202
   442	Crude rubber (not reclaimed)	4208
   520	Copper: Ingot, matte, and pig	5214
   521	Copper, brass, and bronze: bar, sheet, and pipe	5215
   522	Lead and zinc: Ingot, pig, or slab	5216
   523	Aluminum: Ingot, pig, or slab	5212
   660	Alcohol, denatured or wood	5238
   661	Sulpheric acid	5237
   662	Explosives, N.O.S.	5239

7. Starting in 1947, Rubber went from being a "Product of Forest" (442) to "Product of Manufacture" (523)
   So be aware of this if you compare the group totals for those categories across this date.

8. "Database code" - Arbitrary code number I assigned to make database easier to deal with (as the same ICC "commodity number" was sometimes used differently over the years, e.g. commodity "20" was "Coke" for 1914-1919, "Hogs" for 1920 - 1927, "Corn" for 1928-1946, and not used after 1946)

9. Note - if you compare the data from the "Annual report on the statistics of railways in the United States to the Interstate Commerce Commission" (source "1" in the list at the top of this page) and "Freight Commodity Statistics, Class I Steam Railways in the United States for the Year ended December 31, [year]" (source "2" in the list at the top of this page), there are small discrepancies in the numbers listed. For all years that the data was reported both places, I used the data from the "Freight Commodity Statistics..." (source "2") document.

Changes in Data Collected for Each Commodity - National/Regional Statistics

Changes in Data Collected for Each Commodity - Individual Railroad Statistics

Notes:

1. Note how there are minor variations on what is reported for national/regional data vs. individual RR data.
   In particular, NO "carloads" data was reported for individual railroads for 1920 - 1923 and 1928 - 1934.
   And no "carloads" data was collected AT ALL for 1917 - 1919

Process

The process I used to transcribe the data is covered TODO: here.

Database Data

In order to provide the ability to search on the different "direction of traffic" figures across the different years, I have added some derived/calculated fields to the original ICC data. Any data in the original ICC reports is in black colored text in the table below.

Some of the added data (shown in blue text) are simple additions/subtractions/divisions of the data in the report, and should be reasonably valid numbers.

Some of the added data (shown in red text) are calculations based on "aggregate" numbers (e.g. tons per carload for the entire U.S.) and hence should be viewed with some suspicion. In other words, they are a "SWAG" (Scientific Wild-Ass Guess).
I include them anyway to allow some comparison of data across longer time spans (and I figure others will create this data anyhow).

Then there's the carload data for 1917 - 1919. This is VERY problematic data (even worse than a SWAG), so I show it in "fuchsia" in the table below.
There no national / regional carload data to start with. Extrapolating using the 1920 data (although this is what I used) is perilous, as this time period is the end of the transition in freight car design from the "all wood" designs of the late 1800s, thru the composite designs of the early 1900s, to the "all steel" designs starting to show up during this period. So I suspect the average "tons/car" is changing rapidly for this period (and probably changing differently for different commodities). So use at your own risk. Your milage may vary. Caveat Emptor.
While I can be justly accused of "Making the data fit the database design" I feel it is still (somewhat) useful, and (again) I figure someone else will do it anyhow, so I might as well do it myself.
If someone can find better numbers for carload totals or (average tons/car) for each commodity for this time period, I would be more than glad to incorporate it into the database.

Database Data for Each Commodity

1917 - 1919
Carloads Originated (Note 1)	Tonnage originated on road			Carloads Received (Note 2)	Tons Received (Note 2)			Total Carloads (Note 3)	Total tonnage		United States average tons per carload (Note 4)	Eastern District average tons per carload (Note 5)	Total Class I carriers		Class I carriers - Eastern District
													Carloads (Note 6)	Total freight tonnage	Carloads (Note 7)	Total freight tonnage
1920 - 1923
Revenue Freight Originated				Revenue Freight Received				Total Revenue Freight Carried			United States average tons per carload (Note 11)	Eastern District average tons per carload (Note 12)	US - All districts Revenue freight originated		Eastern District Revenue freight originated
Carloads (Note 8)	Tonnage originated on road			Carloads (Note 9)	Tons (Note 9)			Carloads (Note 10)	Total tonnage				Number of Carloads	Number of tons	Number of Carloads	Number of tons
1924
Revenue Freight Originated				Revenue Freight Received				Total Revenue Freight Carried			United States average tons per carload (Note 11)	Eastern District average tons per carload (Note 12)	US - All districts Revenue freight originated		Eastern District Revenue freight originated
Carloads	Tons			Carloads (Note 9)	Tons (Note 9)			Carloads	Tons				Number of Carloads	Number of tons	Number of Carloads	Number of tons
1925 - 1927
Revenue Freight Originated				Revenue Freight Received				Total Revenue Freight Carried			United States average tons per carload (Note 11)	New England region average tons per carload (Note 13)	US - All districts Revenue freight originated		New England region Revenue freight originated
Carloads	Tons			Carloads (Note 9)	Tons (Note 9)			Carloads	Tons				Number of Carloads	Number of tons	Number of Carloads	Number of tons
1928 - 1934
Number of tons of revenue freight								Totals		Freight revenue (dollars)	United States average tons per carload (Note 11)	New England region average tons per carload (Note 13)	US - All districts Revenue freight originated		New England region Revenue freight originated
Originating on respondent's line				Received from connecting carriers				Carloads (Note 10)	Total tons carried
Carloads (Note 14)	Terminating on respondent's line	Carloads (Note 15)	Delivered to connecting carriers	Carloads (Note 16)	Terminating on respondent's line	Carloads (Note 17)	Delivered to connecting carriers						Number of Carloads	Number of tons	Number of Carloads	Number of tons
1935 - 1946
Number of tons of revenue freight								Totals		Freight revenue (dollars)	United States average tons per carload (Note 11)	New England region average tons per carload (Note 13)	US - All districts Revenue freight originated		New England region Revenue freight originated
Originating on respondent's line				Received from connecting carriers				Carloads	Total tons carried
Carloads (Note 18)	Terminating on respondent's line	Carloads (Note 19)	Delivered to connecting carriers	Carloads (Note 20)	Terminating on respondent's line	Carloads (Note 21)	Delivered to connecting carriers						Number of Carloads	Number of tons	Number of Carloads	Number of tons
1947 - 1963
Number of tons of revenue freight								Total revenue freight carried		Gross freight revenue (dollars)	United States average tons per carload (Note 11)	New England region average tons per carload (Note 13)	US - All districts Revenue freight originated		New England region Revenue freight originated
Originating on respondent's line				Received from connecting carriers				Carloads	Total tons carried
Carloads (Note 18)	Terminating on respondent's line	Carloads (Note 19)	Delivered to connecting carriers	Carloads (Note 20)	Terminating on respondent's line	Carloads (Note 21)	Delivered to connecting carriers						Number of Carloads	Number of tons	Number of Carloads	Number of tons

Notes:

1. As no other data appears to exist, 1917-1919 carloads originated calculated by dividing "Tons Originated on road" by [1920!] "Eastern district average tons per carload"
   I am hoping the Eastern district data more closely reflects loading on the B&M than does the national data.

2. 1917-1919 Carloads [or tons] received = Total Carloads [or tons] - Carloads [or tons] originated (carloads being based on 1920 data).

3. 1917-1919 total carloads calculated by dividing "Total tonnage" by [1920!] "United States average tons per carload"
   As cars could be coming from any railroad, I am hoping the national average is a better estimate of this number.

4. 1917-1919 "United States average tons per carload" is using the 1920 values.

5. 1917-1919 "Eastern district average tons per carload" is using the 1920 values.

6. 1917-1919 "Total Class 1" carloads back-calculated from 1920 "United States average tons per carload" and 1917-1919 "Total freight tonnage" for Class 1 carriers.

7. 1917-1919 Eastern District carloads back-calculated from 1920 "Eastern district average tons per carload" and 1917-1919 "Total freight tonnage" for Eastern District Class 1 carriers.

8. 1920-1923 originated carloads calculated by dividing "Tons Originated on road" by "Eastern district average tons per carload"
   I am hoping the Eastern district data more closely reflects loading on the B&M than does the national data.

9. 1920-1927 Carloads [or tons] received = Total Carloads [or tons] - Carloads [or tons] originated.

10. 1920-1923, also 1928-1934, total carloads calculated by dividing "Total tonnage" by "United States average tons per carload"
    As cars could be coming from any railroad, I am hoping the national average is a better estimate of this number.
    Note this number is a bit dubious. I tested this calculation for a few years from 1935 - 1963 (when the actual number is known) and, while most differences were smaller, I did see differences of up to ±100% (usually the largest errors were when there were only 1 or 2 cars, but none the less they are there)

11. 1920-1963 "United States average tons per carload" calculated by taking United States "Revenue freight originated" data "Number of tons" divided by "Number of carloads"
    I have done some spot checking and this calculation matches the data published in the ICC reports (the average tons per car data was published in the reports from 1930 onwards, albeit the published data is to 2 decimal places, and I used the full precision available in Excel).

12. 1920-1924 "Eastern district average tons per carload" calculated by taking Eastern District "Revenue freight originated" data "Number of tons" divided by "Number of carloads"

13. 1925-1963 "New England region average tons per carload" calculated by taking New England region "Revenue freight originated" data "Number of tons" divided by "Number of carloads"
    I have done some spot checking and this calculation matches the data published in the ICC reports (the average tons per car data was published in the reports from 1930 onwards, albeit the published data is to 2 decimal places, and I used the full precision available in Excel).

14. 1928-1934 Carloads "Originated and Terminated" calculated by dividing "Tons Originated and Terminated" by "New England average tons per carload"
    I am hoping the New England region data more closely reflects loading on the B&M than does the national data.

15. 1928-1934 Carloads "Originated and Delivered" calculated by dividing "Tons Originated and Delivered" by "New England average tons per carload"
    I am hoping the New England region data more closely reflects loading on the B&M than does the national data.

16. 1928-1934 Carloads "Received and Terminated" calculated by dividing "Tons Received and Terminated" by "United States average tons per carload"
    As cars could be coming from any railroad, I am hoping the national average is a better estimate of this number.

17. 1928-1934 Carloads "Received and Delivered" calculated by dividing "Tons Received and Delivered" by "United States average tons per carload"
    As cars could be coming from any railroad, I am hoping the national average is a better estimate of this number.

18. 1935-1963 Carloads "Originated and Terminated" calculated in a two step process.
    Calculated by dividing the "Originated and Terminated Tons" by the "New England average tons per carload"

19. 1935-1963 Carloads "Originated and Delivered".
    Calculated by dividing the "Originated and Delivered Tons" by the "New England average tons per carload"

20. 1935-1963 Carloads "Received and Terminated".
    Calculated by dividing the "Received and Terminated Tons" by the "United States tons per carload"

21. 1935-1963 Carloads "Received and Delivered".
    Calculated by dividing the "Received and Delivered Tons" by the "United States tons per carload"
22. Finally...Sometimes the carload count (based on an average tons/car) ended up 0 (less than .5) even though tons were greater than 0.
    So if that happens (if tons were more than zero, but carloads calculated to 0) I arbitrarily set the carload count to 1.

Database Tables

• Table 1 - Commodity Name

  Field	Definition / options	Notes
  Unique ID	Integer	Arbitrary number database creates when record added to the database
  Year_Min	Integer	First year this definition is valid
  Year_Max	Integer	Last year this definition is valid
  Database_Code	Integer	Unique code
  ICC Commodity Code	Integer	Code number used by ICC ("-1" if none ever assigned)
  ICC Commodity Code Suffix	1 character	For "A" of "701A"
  ICC Commodity Descr.	Text	ICC descriptive text

• Table 2 - Scope Names

  Field	Definition / options	Notes
  Unique ID	Integer	Arbitrary number database creates when record added to the database
  Scope Index	Integer	1 = National scope code 2-4 = District scope codes 10-17 = Region scope codes 100 - ? = Railroad scope codes
  Name	Text	"National", or name of district / region / railroad

• Table 3 - Traffic

  Field	Definition / options	Notes
  Unique Index	INT (primary key)	Assigned when data loaded into database
  Year	SMALLINT
  Database Code	SMALLINT
  Scope	TINYINT	One of following list:
  	4	National
  	3	District
  	2	Region
  	1	Railroad
  Scope Index	Integer	Scope Code (see above)
  Commodity Source	TINYINT	One of following list:
  	1	Listed in ICC report
  	2	"Supercode" (derived by adding ICC data together)
  	3	Not Available (1942-1943, specific commodities)
  Traffic Type	TINYINT	One of following list:
  	2	Originated
  	3	Received
  	4	Originated and Terminated On-Line
  	5	Originated and Delivered to Connecting Carrier
  	6	Received and Terminated On-Line
  	7	Received and Delivered to Connecting Carrier
  	1	Total
  Cars Source	TINYINT	One of following list: (from most to least reliable)
  	1	Data published in ICC reports.
  	2	Direct calculation based on published data
  	3	Indirect calculation based on average data
  	4	Extrapolated data
  Tons Source	TINYINT	One of same list as "Cars Source"
  No. Cars	INT UNSIGNED
  No. Tons	INT UNSIGNED
  Percent Total Carloads	DECIMAL(5,2)	Carloads for this commodity / carloads for Database Index 6100
  Percent Total Tons	DECIMAL(5,2)	Tons for this commodity / tons for Database Index 7001
  Tons per Car	DECIMAL(5,1)	Number of tons / number of cars

  Other indices based on [Year + Database code] and [Database code]

• Table 4 - Revenues

  Field	Definition / options	Notes
  Unique Index	INT (primary key)	Assigned when data loaded into database
  Year	SMALLINT
  Database Code	SMALLINT
  Scope	TINYINT	Railroad Future: National, District or Region?
  Scope Index	Integer	Scope Code (see above)
  Commodity Source	TINYINT	One of list above
  Revenue	INT UNSIGNED
  Percent Total Revenue	DECIMAL(5,1)	Revenue for this commodity / revenue for Database Index 7001

  Other indices based on [Year + Database code] and [Database code]

Common Abbreviations

• N.O.S. = "Not Otherwise Specified"
• C.L. = Carload
• L.C.L = Less than Car Load
• N/A - Not available
• K.D. = Knocked Down (?)

Copyrights

As the original data is government data, I do not think it is copyrighted.
If you want to use the data for non-commercial purposes, feel free to do so, but I would appreciate some recognition for the effort of converting "ink on paper" to numbers that can be analyzed.
Note, on the documents I used, the following years include the legend "Government Printing Office":

• 1925 - 1939 inclusive
• 1944 - 1947 inclusive
• 1949
• 1952 - 1955 inclusive
• For 1956 - 1957, in the copies I was using, the cover page that normally had the "GPO" legend on the back was glued to a manila cover, so was not able to be examined.
• 1958 - 1959 inclusive
• 1961 - 1963 inclusive

This webpage is copyright 2011, B&MRRHS, to the extent permissible by law.

History of this web page

Notes to myself

• MySQL
• MySQL Query
• Pentaho