edit in Google

This dataset shows logs from naval ship U.S.S Vermont over the course of almost one year, from January 30th 1862 to January 23th 1863.[1] The U.S.S Vermont was built in 1818 at the Boston Naval Yard in Boston, Massachusetts and served as a store and receiving ship for almost 100 years, to be condemned and sold on April 17th, 1902.[2] The ship's log shows navigational and weather information for each hour of every day as well as general remarks for each day. This log is formatted into a hardcover book. On each page, a table is printed on the left hand side. Horizontally is the relevant information for each hour, showing Knots which indicates the ships speed, fathoms which indicates the water depth, courses which indicates in what direction the ship is traveling, wind direction and speed, weather conditions, air and water temperature, and barometer which indicates pressure. Between noon and 1PM, there is space to include the distance traveled per log, latitude, longitude, current, and variation which from what I understand refers to magnetic variation which is used to derive the bearing of a landmark with respect to true north. The right hand side of the page shows the log date, as well as space for the crew to input general remarks for the day. It is clear in this dataset, that over the course of the year, several crew members were responsible for inputting data, as the handwriting seemed to change throughout the course of the log.

Looking into the history of the ship, I found that it was not commissioned until January 30th, 1862, due to space and fire safety concerns, which means this ship was commissioned the same day this log began.[3] I found that on February 24th 1862, the same night the ship left port for the first time as a commissioned naval ship, a violent storm off the coast of Massachucetts tore away her sails, lifeboats, and rudder, flooding the ship and destroying most of the interior. The violent storm though, did not stop the log from being filled in. The date February 24th, 1862 took up two pages on the log. Both pages seem more damaged than most, indicating perhaps, the severity of the storm. The data logger writes, “During the night the ship made great quantities of water on the berth deck,” indicating that the area of the ship where the crew slept was flooded. The log continues, “so much so as to wash mess chests,” indicating that their food supplies had been swept away by the storm. The log dated February 26th, 1862, shows that the ship lost its rudder and had been drifting aimlessly for some time. I can also read that at 11PM, the U.S.S. Vermont made a distress call to a nearby ship, sending a man on board to persuade the captain to return to port and relay the condition of the vessel to naval authorities. On March 7th, it was indicated that rescue crews came aboard and helped to repair the ship. It was interesting to find that throughout this storm, wind and weather records were filled in almost hourly, though not in their correct positions. There was no data filled in regarding the ship's speed and positioning. By March 10th 1862, we begin to see speed data, indicating that the ship had set sail again. If this dataset were to be digitized, it would be difficult to interpret and account for this event. If the digitized data only included information inputted into the table, we would assume the ship set sail, stopped for a while, and then resumed sailing. We would know nothing about the storm which nearly put this ship at the bottom of the ocean. This is extremely important because it helps us see the shortcomings of data digitization efforts and what kinds of information will potentially never be uncovered, lost in the process of data digitization. These records and the events that transpired between February 24th 1862 and March 10th 1862 prove to be a prime example of the types of information that would naturally be lost in data digitization efforts.

What I would like to focus on here are the types of data that can be lost to digitization and their consequences. A similar document of this era would be the U.S. Census, which was conducted using similar tables, printed on a piece of paper and filled in by hand. By intentionally recording insufficient information of Native populations and their territories, the US government shaped a narrative that these areas of land are open and primed for westward expansion. In Charting Progress: Francis Amasa Walker’s Statistical Atlas of the United States and Narratives of Western Expansion, Thomas P. Kinnahan writes that there were “verbal and visual frameworks that partially elide the Native American presence in the West, celebrate the emerging industrial order, and ultimately draw upon what I term a rhetoric of the economic sublime to reconfigure the West as a resource frontier rather than a field for agrarian settlement.”[4] This differs greatly from our U.S.S Vermont ships logs, but it helps to paint the picture that when data is left out, intentionally or unintentionally, data digitization can create an inaccurate or imperfect rendition of history. Reading between the lines and understanding what has been left out and why, helps us to piece together the puzzle and understand the data more thoroughly. Examining other data sets such as Lewis and Clark’s Journal logs[5] imply some of the same issues when it comes to data digitization. The journal logs include a handwritten table with cells for daily weather, wind direction, and rainfall. Similarly to the U.S.S. Vermont ship logs, on the right hand side of the page is an area for remarks. If this dataset was digitized and only included information from the table, information from their remarks may never be discovered.

If we were to digitize the data from the U.S.S. Vermont, I would implement a plan that would account for both the data included within the table, and also the crew’s daily remarks. The handwriting is often hard to read and occasionally illegible, so I would suggest a section that summarizes generally this information. This could be a dangerous decision, as it gives discretion to the digitizer on what information from the remarks they wish to include or omit, but I think it would be generally too time consuming and disjointed to include every word of these remarks. This data set contains 188 total scans with 374 potential days worth of information. Since this log only contains data from 358 days of sailing, it is clear that some pages are duplicated. Each table contains 223 cells. I believe that each table could take on average 20 minutes to digitize. For the remarks category, I believe that it could be read and summarized in around an hour. Combining the table and remarks digitization, multiplied by 358, this dataset should be able to be digitized in 477 hours and 20 minutes. The data is published by the National Archives and Records Administration (NARA), which includes an area for users to transcribe data themselves, though I am not able to see how much of this dataset has been transcribed. In addition, there is only a transcription feature included, with no area to input values into a table, likely rendering this feature useless to this digitization effort. With data from the observed daily longitude and latitude, we could potentially visualize the ship's course over time, and quite possibly where the storm hit, enabling divers to uncover lost wreckage such as the ship's sunken rudder. In addition, we could visualize the courses the ship took most often, and which ports it frequented. We could use a heat point map to see visually the areas where the ship was most often with density lines in between to signify the routes taken most often between these points. All of this would depend on the accuracy of their longitude and latitude readings, as well as the frequency they were inputted.