Feeds:
Posts
Comments

[Cross-Posted at Digital Harrisburg]

As students have noted recently (here, here, and here), everyone feels a sense of accomplishment and relief in bringing to completion the “City  Social” U.S. Census data project for the digital history class. What students have been able to accomplish in the last six weeks is truly remarkable both on an historical and logistical level. We’ve put some 60% of the population of Harrisburg in 1900 “on the map” and done so within the context of a couple of coterminous classes (Digital History and GIS) using over-the-counter software (MS Excel, Access, and ArcGIS) and scans of the original census records available via Ancestry. Every student in my digital history class spent about 20 hours keying this data, and students in GIS courses at Messiah College and Harrisburg University have put in their share of hours.

On the data entry side of the project, the records keyed by students passed through multiple checks for data consistency and integrity, which required fussy, tedious, mind-numbing scanning. Each student had the assignment of looking at another student’s 2,000 records to check and standardize the data. Here’s the workflow we adopted to move data from rough draft to final version:

Step 1: General Check for Completeness

This was a ‘big picture’ check to make sure all of the records were present and accounted for – to rule out the potentially cataclysmic possibility that a student assigned to digitize the records from, say, Ward 2, District 2, in the year 1900, accidentally keyed Ward 2, District 3 from 1920. In this step, students were required to:

  • Check ward, district, and sheet numbers on the 10 individual worksheets within each spreadsheet. Those numbers should agree with the file name of the excel spreadsheet and number sequentially from 1-10.

  • Check order and number of columns. There should be 25 columns (A-Y in Excel) representing 25 different census values, and follow the order we agreed on in class.

  • Check number of names per worksheet. There should be 10 worksheets each with 100 names. If not, double check against the original census sheets.

  • Scroll through each worksheet quickly to see if any data was left out

  • Where street names and/or numbers are absent, highlight cells in yellow

If major problems, note them in a document on Google Drive called “Troubleshooting” and return spreadsheet to the author for correction.

Step 2: Careful Data Check

After this general data check, I asked students to look at their classmate’s data more carefully:

  • Check the first name and last name on each worksheet against record #1 and 100 on the original census sheet

  • Check street name on each worksheet against the original census sheet

  • Run a full check on the first ten (1-10) and last ten (91-100) records on each sheet compared with the original census records. This constitutes a 20% quality check on all the data. Are there any mistakes? If so, highlight any cells with errors in blue and note the mistakes in the Troubleshooting document

This 20% quality check offered us a general sense of the data quality for each student, and how carefully we needed to recheck particular districts. In cases where multiple mistakes were found in this sample data check, we will run a more thorough quality check on 100% of the data.

Step 3: Aggregate and Normalize Data

At this step, I took over. I downloaded all spreadsheets from Google Drive to my computer and created a “Master” worksheet for every student combining individual census sheets into a continuous series of 1,000 names for the specific enumeration district. Scrolling through sets of 1,000 names, I was able to check for additional mistakes, missing data, or inconsistencies.

I then created a master Ward spreadsheet for every ward that aggregated all the names keyed for each ward. So, Ward 1 data keyed by one student was combined with Ward 1 data keyed by other students in a file called Ward1_Master. There were ten master spreadsheets, each with thousands of records.

To check data quality, I tried importing the records into Access. Because Access requires that you to categorize fields as text, number, etc.., it generates errors when one kind of data from Excel (e.g., text) doesn’t agree with the field’s properties (e.g., number). So, for example, an address value of “109” imports perfectly from Excel to Access when address is defined as “double” (number) in Access. But if the cell’s value in Excel has text (“BLANK”) instead of numbers, the import generates an error message. Doing this with data from each ward allowed me to highlight additional problem spots in the data entry, which I was then able to correct in the Excel Ward Master files.

Step 4: Create Unique IDs

At this stage, I created a Unique ID for the records in each ward. As I noted in an earlier post, every individual census record has to be tied to a unique identifying number that is machine-friendly and recognizable by ArcGIS. After considerable deliberation with Professors Jeff Erikson and Albert Sarvis about what this ID should look like, we decided that the best way to connect census data with GIS shape files was to create a connecting ID via the street address – with address number followed (without space) by the upper case spelling of the street name.

In the image of the spreadsheet below, the ID is shown in column G. The Spigelmyer family, for example, shown in Rows 4-6, lived at 1106 Ninth Street in the first ward. Their unique ID was 1106NINTH. That same ID is also given to 1106 Ninth St in the GIS. To create the unique ID in the census record, I copied column D into column E, and then removed the suffix “Street”, “Avenue”, “Lane”, etc… using Find / Replace. Then, I capitalized the street names in column F using the UPPER function, and combined the upper-case street name with street address (in column H) using the CONCATENATE function. The result is the ID you see in Column G.

Ward-Master

When the two values match, the GIS will correctly import the full set of data from the census record. The unfortunate reality is that the two do not always match. Students working from census data used the inconsistent spellings of street names recorded by the census takers, while students working on GIS maps used the spelling of street names preserved on the 1901 Harrisburg Atlas.  A census recorded labeled 101HANNAH will not relate to a shape file named 101HANNA – even though the two represent the same place on the map. Correcting these inconsistencies will require a good deal more fussy work.

Results

The result of all of this work is a Microsoft Access database with 28,397 names, and dozens of GIS files. The following table shows the number of records keyed from each ward compared to the total number of individuals listed in each ward (the total number is actually an estimate of total number). As you can see, students have keyed 100% of Wards 1, 3, 4, and 10, about 70% of Wards 2 and 8, and about 25-50% of Wards 5,6, and 9. We can talk with total confidence about the social make-up of certain wards of the city and can talk reservedly about the rest.

Ward Number Keyed Number Total Percentage
1 3909 3909 100.00%
2 3000 4269 70.27%
3 2144 2144 100.00%
4 2644 2644 100.00%
5 2000 4313 46.37%
6 3000 12869 23.31%
7 3000 7451 40.26%
8 3000 4435 67.64%
9 3000 6456 46.47%
10 1902 1902 100.00%

I am hoping we’ll have some concrete visuals to show in the next few weeks, and that my students will continue to discuss the census data as they have been this week and last. I’ll present some tabulation and analysis at some point as well. But we have another project to move forward: City Beautiful. Expect much more discussion now about that.

All of this marks only a start, of course. We’ll need all of next fall semester to finish the data entry for Harrisburg’s population in 1900, normalize, run integrity checks on the census records, and fix the wide range of inconsistencies between data entry and GIS. Fortunately, several students will continue working on this project five to ten hours per week in the fall.

Students who have contributed to this project via GIS and history classes at Messiah and Harrisburg University deserve this moment of accomplishment. In our first semester working on this project, we’ve put Harrisburg’s population in 1900 literally on the map. We’re looking forward to what this can now tell us about the city social and beautiful at the turn of the last century.

Last Thursday, a group of history club students teamed up with the Oakes Museum of Natural History to work toward preserving a dilapidated cemetery of an historic Brethren community of the late 18th to early 19th century. The Little Bermudian Brethren community was a plant, or off-shoot, of the religious community at Ephrata Cloister in the late 1700s. Migrating across the Susquehanna River from Lancaster county, this group of Swiss German Christians settled in scattered farms in what is today Washington and Franklin township in northern York County. We’ve written about one of the leading members, Abraham Stouffer, at this site. His friends and family are buried in the Little Bermudian Cemetery (a.k.a. Asper Burial Ground).

As I noted here about our fieldwork on Service Day, our plans this year included:

1. Excavating an embankment that we suspect contains bulldozed gravestones

2. Documenting gravestones

3. Completing a Pennsylvania Bureau of Historic Preservation Cemetery Survey Form to register the site with the state

4. Conducting coring tests within the cemetery in areas we suspect have buried tombstones

5. Weeding the cemetery with gardening tools

6. Taking photographs

The project was even more successful than I imagined it would be. Our group of seven students, plus Ken Mark (director of the Oakes Museum), Diane Phillips (owner of the Stouffer Farm), and myself, were able to accomplish quite a lot on a beautiful spring day, and made some important discoveries in the process.

First of all, we continued to excavate the embankment that we had started digging in Service Day 2012. In that year, we had discovered a footstone fragment and a couple of other stone objects that clearly derived from markers in the cemetery. These confirmed a written report dating to 1941 that plowing had destroyed some of the cemetery and broken up the stones. Fortunately, this year, excavation in two trenches 2 x 2 m in size produced no clear examples of additional stones. There was, in short, only a little evidence from the embankments for this report – and that, I think, is a good thing, in that our sample rules out an extensive area of displaced grave stones.

Below, Greg Slye in Trench 1 and Tyler Stone in Trench 2 at the start of the day.

2014-04-10 10.57.59_m2014-04-10 10.16.22_m

Katy Kaslow digs into Trench 2

2014-04-10 10.58.08_m

In the cemetery itself, we used an auger to locate buried head and footstones in the cemetery, and conducted a careful inspection of all the stones in the cemetery. Our goal was to determine if we had located every known individual of the cemetery. A family of groundhogs had displaced enormous amounts of soil over many decades and buried and moved stones. There is a local legend that an area resident had once removed stones from the cemetery to his house until local outrage forced him to return them. We were interested, then, in whether the stones we see today are still marking the place of original burials, or whether they represent stones out of their original context.

To our delight, the use of the auger, combined with scraping away the excess deposits of groundhog deposits, revealed the bottom of a headstone in German. The inscription belongs to the stone of Rael Beisel, the wife of Peter Beisel, who was the nephew of the famous Conrad Beissel of Ephrata Cloister. Although Rael’s name is now missing from this stone, Harry Sinner’s inventory of the cemetery in 1941 recorded the same German phrase we see here: Im Jahr 1794, der monatden Hornung dn 1. In other words, we had unknowingly uncovered part of a headstone of a missing community member that was known to observers of this cemetery 70 years ago.

2014-04-10 10.59.06_m

In a similar way, we documented an entirely unknown member of the community. One can see from the fragment below the letters AIAHB. Our current interpretation is that this represents the headstone of an ISAIAH B-. Who this Isaiah is we will have to determine. But Harry Sinner does not note the stone in his record in 1941.

2014-04-10 14.57.39_m

Using the auger, we also located the footstones of John Pentz and Philipena Pentz, labeled respectively JP and PP. These was important finds because they proved – in their spatial relationship to the headstones – that the they marked the exact place of burial. It rules out the theory at least that these stones have moved around significantly.

Katy Kaslow, Kelly Henderson, and Ken Mark act as living scales below.

2014-04-10 11.27.16_m

In a similar way, we used the auger to locate the footstone of Abraham Stouffer himself. Marked by the letters A. S., this 75 centimeter stone was entirely buried through groundhog displacement. This is the top of the stone just below surface level.

2014-04-10 11.29.18_m

Below, Tyler Stone, Sean Barron, and Katy Kaslow uncover the footstone.

2014-04-10 14.20.43_m

After significant effort, most of the stone had been revealed.

2014-04-10 14.40.34_m

Tyler and Sean show off the footstone that had been completed buried.

2014-04-10 14.54.07_m

We re-erected this footstone at the proper place above ground on the east end of the burial of Abraham Stouffer. Below, the group poses in front of the stone. From left to right: Greg Slye, Shane Reed, Kyle Polinka, Tyler Stone, Katy Kaslow, Sean Barron, and Kelly Henderson

2014-04-10 15.23.36_m

We will probably return to the cemetery at least one more season – in the Historical Archaeology class in the fall – to straighten the stones, remove excess groundhog piles, and seek additional buried foot or head stones. The results of our geophysical survey here in 2013 will guide our efforts.

Thanks to all who came out and made this another successful Service Day event.

Since we’re in the midst of planning schedules for the fall, students may want to think ahead about the spring semester.

I queried the other historians in the department about what upper-level history courses that will be offered in the spring, and here is what we came up with:

HIST 302. Ancient Rome

HIST 304. Tudor-Stuart England: 1400-1700

HIST 319. Topics: Archaeology and History of Cyprus (May-term)

HIST 345. Pennsylvania History

HIST 355. U.S. Urban History

HIST 3xx. Non-Western History (Topic TBA)

For full course descriptions, visit this page.

*The May-term course to Cyprus is now accepting applications. See this link for more information.

**Note that the next 300-level Modern Europe requirement for History majors will be taught in Spring 2016. HSST students who need this requirement for graduation in 2016 should enroll in HIST 322 next Fall since you will be student teaching in Spring 2016.

Messiah College’s official induction into Phi Alpha Theta (History Honor Society) will be on Wednesday, April 30, 2014 at 4:00 pm in the Boyer Atrium. A reception will follow the induction ceremony. All friends of History are encouraged to attend and welcome the new class.

Please contact your adviser if you have any questions about joining the History Honor Society. Membership is for life. Life-time dues in this International Honor Society are $40 (pay Messiah College’s business office Account # 2051-6180 "Dues and Subscriptions"). Please keep your receipt from the Business Office. Members will receive a one-year subscription to "The Historian" and the Phi Alpha Theta newsletter. For more information about Phi Alpha Theta go to http://phialphatheta.org/ and for a membership form go to http://www.phialphatheta.org/data/memappform.pdf (or print the attached PDF file). Membership applications are due to your adviser by Tuesday, April 15th.
Membership requirements are:  Undergraduate students must complete at least 12 semester hours in History (4 courses) with a GPA of at least 3.1 in History and must have an overall GPA of 3.0 or better. Membership is not limited to History majors.

With registration now open, here’s one more upper-division history class for fall: Dr. Huffman’s “Medieval Europe” (HIST 310). Here’s a primer:

The course offers a survey of the cultural, social, economic, religious, and political developments in Europe from the eighth to fifteenth centuries A.D. Major themes include the emergence of medieval social institutions and modes of thought, Christian monasticism and spirituality, and the cultural interactions of the Latin West, the Byzantine East, and the Islamic world.

The course will entail the emergence of "the West" from its Carolingian foundations through eras punctuated by prosperity, rapid growth, reform, and renewal as well as by warfare, pestilence, and dislocation.  In all this there will be an emphasis on (a) primary-source analysis of classic texts of medieval spirituality and (b) historiographical debates about the meanings of the Middle Ages for peoples and nations during the modern era as well as into the 21st century.

Would you like to join history majors from other Christian colleges who will be presenting research papers at the biennial Conference on Faith and History undergraduate conference?  The conference takes place at Pepperdine University in Malibu on September 24 and 25, 2015.   Messiah College history majors have presented at this conference in the past and have found it to be a very rewarding experience.

If you have a paper (Historical Methods or a 300-level course) you would like to present, or if you want to know more about the conference, please stop by and see Dr. Fea as soon as possible. (He will be attending the conference).  The deadline for proposals is April 15.  Some financial aid may be available for the trip.

Most of our students are probably aware now of the courses that we will be offering in fall semester, but here’s the list anyway. We will continue to post more detailed summaries of these courses in the next week or so. We may also post a list of our projected courses for Spring 2015. Stay tuned.

 

HIST 099: Introduction to History (1 credit). Fea.

HIST 101: Western Civilization, before 1500. Gen Ed: European History. TBA.

HIST 102: Western Civilization, 1500 to Present. Gen Ed: European Hist. Wilson.

HIST 141: U.S. History Survey to 1865. Gen Ed: U.S. History. Fea.

HIST 142: U.S. History Survey since 1865. Gen Ed: U.S. History. LaGrand.

HIST 172: World Civilizations, 1500 to Present. Gen Ed: Non-Western St. TBA.

HIST 258: Historical Methods. LaGrand

HIST 305: Archaeology and Historical Interpretation. Pettegrew

HIST 310: Medieval Europe, AD 700-1500. Huffman.

HIST 322: Modern Europe, 1789-1918. Wilson.

HIST 344: Civil War America. Gen Ed: U.S. History. Jordan.

HIST 399: Topics: The History of American Evangelicalism. Fea.

HIST 399: Non-Western History Course. TBA.

HIST 401: Historiography and Philosophy of History. Wilson.

LATN 101: Fundamentals of Latin I. Gen Ed: Languages and Cultures. Pettegrew.

LATN 201: Intermediate Latin. Gen Ed: Languages and Cultures. Huffman.

Follow

Get every new post delivered to your Inbox.

Join 249 other followers

%d bloggers like this: